TheChek News Flow

Jigar Shah, who runs the Energy Department’s loan program, is trying to hand out a lot of money for green-technology projects, while navigating an unforgiving political environment

Jigar Shah is living an investor’s dream, one with more strings attached than a symphony orchestra.

Shah has $400 billion of government funds to pour into businesses touting green-energy projects. But he has to do it under the eye of critical lawmakers, cautious bureaucrats and the White House, which has already clashed with him on the politics of his lending juggernaut. Losses are likely and will be frowned on by Congress.

The line for Shah’s cash stretches to 150 companies seeking $127.7 billion in loans, ranging from new companies with unproven products to giants such as General Motors and PG&E, the California utility blamed for deadly wildfires. Funneling that much money to climate startups in a short time would be near impossible. Shah has begun writing bigger checks, including a record $9.2 billion commitment to a Ford joint venture making batteries in Tennessee and Kentucky.

The source of Shah’s financial firepower is the Energy Department’s Loan Programs Office, an overlooked piece of the Biden administration’s strategy to address climate change. Largely quiescent for almost a decade, the office is designed to finance businesses that are important to the country’s energy transition but unable to borrow from traditional lenders, often because their technology is seen as too risky or because the terms are too onerous.

“We would absolutely look at investing alongside them,” said Jehangir Vevaina, a managing partner at Brookfield Asset Management who helps oversee the private firm’s $15 billion energy transition fund. That fund, one of the largest of its kind, typically invests in a company’s equity, which can become less risky when government loans give businesses a stamp of approval, as well as lower borrowing costs than commercial banks.

Climate-related provisions in last year’s Inflation Reduction Act gave Shah’s office a windfall, multiplying its lending capacity 10-fold. That pile of cash is at least 20 times as big as most private green-energy funds, giving Shah and the Loan Programs Office a major role in shaping the American energy landscape.

That is how Shah found himself in early 2021 calling hundreds of clean-energy executives to pitch the loans his office could provide. Primary targets were clean-energy startups that had raised at least $100 million in equity financing. He also wooed big businesses with the resources to pay back large loans.

Some were reluctant to apply, worried about the complicated approval process and the risks of taking a government loan. Shah, eager to get funds out the door, can be impatient. In September, he pressed a startup company that has a plan for recycling batteries to borrow hundreds of millions of dollars from the federal government to construct a plant.

The company’s chief executive, Ajay Kochhar, was hesitant, unsure how quickly it could repay. “Get your ass to Pittsburgh,” where a clean-energy conference was about to start, Shah told the executive, according to people familiar with the conversation.

At a coffee shop soon after, Shah told Kochhar, of Li-Cycle Holdings, that its recycling plant could easily generate enough revenue for repayment. Five months later, the two announced a $375 million federal loan.

Shah’s office is “the clean-energy bank of the United States,” said Peter Davidson, who led it from 2013 to 2015.

With its burst of funding, “the floodgates have really opened,” he said.

The loan program is part of the reason the Inflation Reduction Act’s tax credits and spending represent one of the largest outlays of taxpayer-financed industrial stimulus since the 1930s New Deal.

In early June, Shah’s office committed $850 million to startup battery maker Kore Power. The loan would fund what the company calls the KOREplex, a giant battery manufacturing facility in the desert about 35 miles west of Phoenix.

Shortly after, the Loan Programs Office announced the record commitment to the Ford battery venture. The $9.2 billion agreement is bigger than the $5.9 billion Ford borrowed from the office starting in 2009, when it was struggling through the financial crisis.

Solyndra PTSD

Hanging over the Loan Programs Office’s every move is what Shah called “Solyndra PTSD.” Despite the office’s successes, which include backing Tesla, it remains dogged by a busted loan to solar-panel startup Solyndra.

Solyndra failed in 2011 after China flooded the market with low-price panels. In addition, an investigation by the Energy Department’s inspector general found Solyndra had misrepresented facts and omitted key information in getting the loan. The $535 million loan that went sour made staffers cautious, borrowers nervous and critics of the program aggressive.

Rep. Cathy McMorris Rodgers (R., Wash.), chair of the House Energy and Commerce Committee, has called the expanded funding “Solyndra on steroids” and said the Inflation Reduction Act’s new spending and lending “has heightened the risk for waste, fraud and abuse.”

She said her committee is in regular contact with the loan office.

Shah’s first loan deal after taking over in 2021 was a $1 billion commitment to Monolith, a company that aims to produce hydrogen from natural gas. The company’s technology will capture the carbon that the process yields and turn it into a material for everyday products such as tires.

Clean hydrogen is an alternative to oil and gas.

Several members of Shah’s staff worried the deal was too risky, people familiar with the matter said. Shah argued it was safe because it required Monolith to set aside revenue and meet rigorous requirements before getting the money, such as showing its production process would work at scale. Monolith hasn’t yet met them.

Last summer, Monolith raised more than $300 million in equity from investors including BlackRock and NextEra Energy, the most valuable power company in the U.S.

Staff members raised concerns about a potential loan to a company called Syrah Resources, a producer of graphite, which is used in rechargeable batteries, people familiar with the matter said. Raw material for its Louisiana processing facility would come from a mine in Mozambique, the scene of terrorist attacks, raising concerns that disruptions would threaten Syrah’s project. Shah pushed ahead, on the grounds that the program was protected in the deal and the U.S. needed to lessen its dependence on China for graphite. The office issued a $102 million loan last summer.

When Shah supported granting a loan for an Occidental Petroleum project, White House officials told him it could backfire. The project involved removing carbon from the atmosphere via a new technology, then injecting it into the ground to extract more oil. Administration officials worried about a backlash from environmentalists, a person familiar with the discussion said.

Shah argued it was worth exploring as a way to develop carbon-removal technology. The loan application is still working its way through the office. Occidental’s first big carbon- removal plant is under construction.

Despite having lived for decades around Washington, D.C., Shah seems more like a creature of Silicon Valley’s high-tech culture than a Beltway denizen. He co-founded a solar-energy company, SunEdison, in 2003 with a home-equity line of credit. It revolutionized the way businesses and homeowners paid for rooftop solar panels. Under its financing concept, which Shah drew up for a business-school class project at the University of Maryland, the panels are typically paid for over a 20-year stretch, in part with buyers’ savings from generating their own power, making the panels almost free in the long run.

Installations exploded, and SunEdison became North America’s largest rooftop-solar provider. Shah left in 2008, the company was sold in 2009. Years later it went bankrupt after an aggressive growth strategy backfired.

Shah co-founded clean-energy investment firm Generate Capital about nine years ago. His effusive personality and list of contacts helped build Generate into one of the largest clean-energy investment firms. At Generate’s San Francisco office he would hold court in a cavernous space known as the “Jigar-torium.”

When approached by the Biden transition team about leading the Loan Programs Office, Shah was reluctant. On a podcast he used to co-host called, “The Energy Gang,” he once called the office “irredeemable” because it was doing so little.

Patti Poppe, now the CEO of PG&E, listened to the podcast in the mornings while exercising on her treadmill. “It would make me run faster because he’d make me mad,” she said. Shah often criticized utilities for moving too slowly. Poppe eventually invited Shah to talk to the management team at her previous job in Michigan and became convinced the industry needed to be more aggressive.

At PG&E, she is seeking a roughly $7 billion loan to upgrade and bury the utility’s outdated power lines, to reduce wildfire risk and keep up with rising electricity demand driven by electric vehicles.

‘Damn you!’

Before Shah took the job, his Generate colleagues told him accepting was a dumb idea unless he could make the office more efficient, he said in an interview at his Energy Department office, clad in his blue fleece vest and Stan Smith tennis shoes.

He outlined his demands, including provisions that would make it easier to lend to companies in the electric-vehicle supply chain. On a call with DOE officials, they agreed to all of his conditions, he said.

“I was like, ‘Damn you!’ ” he recalled.

He tripled the agency’s staff to roughly 250 and recruited debt experts from banks. He sought energy specialists such as Bill Magness, a former CEO of the Electric Reliability Council of Texas, that state’s power grid operator.

According to Magness, Shah invited him to meetings even before he agreed to join. “How could you not do it?” Shah told him, Magness said. Magness was a consultant for the office for a year before departing in 2022.

In April, Shah expanded on the rooftop-solar financing model he developed for his first company by improving access to loans for people with below-average credit scores. Through a $3 billion commitment to home-solar company Sunnova, the office would guarantee that even if some users default, many investors would be repaid. Shah is confident defaults will be low, and the backstop won’t be needed.

“If you have a normal government person coming into this spot, they’ll never think of something like that,” said Sunnova’s chief executive, John Berger.

The Loan Programs Office had largely been dormant since the second Obama term. The bulk of the office’s loans in the last decade went to utilities building the Vogtle Electric Generating Plant, a nuclear power project in Georgia.

Shah’s tenure and the program’s aggressive lending could prove short-lived if Republicans win back the White House next year. In Congress, McMorris Rodgers has criticized the loan office’s high funding level and promised greater oversight.

Shah says the government is more protected with today’s deals, through provisions that ensure the government will get some money back even if a borrower fails. The program has beefed up goals companies must meet before receiving funds.

The office’s default rate of 3% is comparable to the performance of loan portfolios of commercial banks, Shah has said. It has made money for the government over its lifetime.

All the loans need a series of approvals from a committee of senior Energy Department staff, as well as Energy Secretary Jennifer Granholm, the White House Office of Management and Budget and the Treasury Department.

“If anyone can crack through some of the red tape, it’s a force of nature like Jigar,” said Scott Jacobs, who co-founded Generate Capital with Shah and one other person. “Yet I’m not sure anyone can get through all of the bureaucracy.”

Millions of US borrowers brace for the return of student debt payments

End of a pandemic-era pause and debt-forgiveness scheme will ripple through the US economy as bills come due

This October, after a three-year break, 27mn Americans with student debt will once again have payments due.

One of them is Jacque Adams, a Dallas public school teacher who owes $103,000 in student debt. “I’m an underpaid teacher,” said Adams, 45. “I have three kids; one is starting on her own college [degree]. I’m going to take these loans to the grave with me.”

The Department of Education paused student loan repayments when the Covid crisis began in March 2020, and extended it nine times in an effort to provide financial relief to US households throughout the pandemic.

But that is now ending, after Congress put a halt to the programme as part of an agreement resolving the debt ceiling stand-off earlier this year. Some economists fear that the resumption of payments, combined with a softening labour market and rising interest rates, could dampen consumer spending and raise balances on other consumer debts, weighing on the broader economy.

In yet another blow to student borrowers, the Supreme Court ruled Friday that US President Joe Biden’s $400bn student loan forgiveness plan, which would have cancelled up to $10,000 of debt for borrowers making up to $125,000 — and up to $20,000 for those who received a Pell Grant, a type of loan for those with the greatest financial need — was unconstitutional.

The programme would have completely eliminated the student debts of 20mn Americans, the White House said.

Biden said on Friday that his administration was exploring other legal pathways to help borrowers.

“I will not stop fighting to deliver relief to borrowers who need it the most,” he said.

Some 43mn Americans have students loans, though 6mn are not paying because they are still in school, 4.6mn have defaulted, 3.1mn have deferred payments, and 1mn are in a grace period before payments begin.

Before the pause, the average student borrower paid between $200 and $299 each month on their loans, according to Federal Reserve data. 

In a Morgan Stanley consumer survey, only 29 per cent of respondents said they would be able to make payments on their student loans without adjusting their spending.

Many borrowers have used their extra cash on things such as investing, saving for retirement, or paying down other debts. Others took it as an opportunity to splurge on experiences they knew they would not be able to afford while paying down their loans, said Zenith Wealth Partners managing partner Chelsea Ransom, a financial planner who primarily advises millennials, many of whom have student debt.

Adams, who has not received a bill or accumulated interest on her loans for the past three years, said the forbearance allowed her to save for a down payment on a home.

“People have been thinking ‘we’ll be fine, we’ll get through this’, but I think once you actually see it coming out of your pay cheque, that’s when the real reality bites and when we do see a shift coming through in terms of consumer attitudes,” said James Knightley, chief international economist at ING. “I think it will be quite a painful story unfortunately, because 43mn Americans is a big chunk of the population.”

Estimates vary on the ultimate impact, but based on calculations from the Bureau of Economic Analysis, a full resumption of related payments would be equivalent to about 0.2 percentage points of disposable income, said Aditya Bhave, senior US economist at Bank of America.

“A lot of people really enjoyed the pause and forgot how much they were paying and never thought that payments would return,” Ransom said. Her clients plan to cut back on dining out, travel, and investing in order to fit their loan payments back into their budgets.

“This is a very sobering moment,” Ransom added.

While consumer balance sheets on the whole are healthy — buoyed by a savings buffer and a robust labour market — strains have begun to emerge on the margins. More Americans are falling behind on payments for credit card debt and auto loans, and that pressure could build once the payments pause ends: “$20,000 is not going to help me,” Adams said.

“There could be knock-on effects to other components of household debt,” said Bhave.

“People who struggled to pay their student loans might end up going delinquent on their credit card loans or their auto loans.”

He estimated that so-called “serious” delinquency rates, which are 90 days late or more, could rise nearly 70 per cent over time as a result.

Should the labour market crack more meaningfully, especially as the Federal Reserve’s efforts to tame inflation by rapidly raising borrowing costs start to bite, vulnerable populations could be pushed further to the edge. Wendy Edelberg, director of the Hamilton Project at the Brookings Institution, said some degree of economic cooling should be welcome, with consumer spending and the labour market still “unsustainably strong”.

Ending student loan relief could help that process in some capacity.

“This actually in a business-cycle management sense goes in the right direction because it is yet another force that would reduce spending,” she said, noting that the “crisis” backdrop that prompted these policies had long passed.

Some borrowers say that they still need help to repay their loans amid rising prices, and that the forbearance alone did not provide enough relief.

Adams said that her recent application for a mortgage had been denied because the lender factored her roughly $500 monthly payment into calculations of her debt to income ratio despite the forbearance, preventing her from purchasing a home. She said she has applied for a federal loan forgiveness programme for public servants, but has yet to be approved. 

Joanna Kearns, 42, of Broward County, Florida, said the end of the pause was “unfair.” She echoed complaints shared by many borrowers that the education department did not properly inform them about student debt before offering them loans as teenagers. 

Kearns, who works as a full-time caregiver for her elderly father as he undergoes cancer treatments, owes roughly $60,000 from two degrees.

“I’m trapped and really deserved this break,” Kearns said. “They trapped us for the rest of our lives over one mistake.”

US streaming service introduced ad-supported subscriptions last year

Netflix will develop increasingly targeted and tailor-made advertising formats to win over marketing bosses as it seeks to boost revenues in its recently launched advertising supported service.

The US streaming service held talks with global advertising executives about new plans at this year’s Cannes Lions festival in the south of France after launching the ad-supported service last year.

Advertising executives said innovations could include “episodic” campaigns that would see a series of different but related sequential ads to consumers, which would avoid the common complaint of consumers of being shown the same ad multiple times when watching a series. These could be shown during related shows — for example, light entertainment — but not others.

Netflix was a prominent presence at the vast festival of advertising in its first year at the event, hosting a party in a branded hotel on the seafront and an event with ex-footballer David Beckham. 

Netflix co-chief executive Greg Peters and Jeremi Gorman, president of worldwide advertising at Netflix, spent the week in France courting advertising bosses and brands, according to multiple advertising executives.

They said that their counterparts at Netflix outlined how the US media giant would introduce sophisticated ways for brands to advertise as part of a long-term strategy to redefine how streamers can use their platforms for commercial purposes. This includes allowing brands to more directly target advertising to consumers in ways not possible on linear TV channels.

One Hollywood marketing executive said that the partnership to use Microsoft’s technology was also likely to come to an end next year, which would allow Netflix greater ability to innovate. 

“They’re building [their own technology] in the background. Once they have their own they’ll do free standing. Microsoft is the interim ad server, but that’ll change when they build their own,” he said.

“The plan for Netflix was just to get to market quickly last year. This is not the final way they’re going to do it. They are going to be very creative. There’s going to be a better, a different experience.”

One advertising boss said Netflix would be able to use its data on customers to create more bespoke marketing, albeit in ways still safeguarding customer data protections.

“They’ll know what you’ve seen,” he said. “So the old days of making episodic work may be back because before you could never guarantee what people have seen already. Now you can write 15 episodes of an advert and guarantee that the viewer will see them in the right order. So that’s really interesting.”

Another advertising chief said Netflix now made more money per user on its advertising plans than in the standard tier. He said brands had been positive on the idea of using Netflix to market their products, although he added that the size of the audience was still small.

“Streaming services thought it would devalue their offer. But actually, it’s massively expanded the reach. Brands are extremely keen to be associated with the Netflix content and now they’re starting to think about much more innovation around the format, much more strategic partnerships.”

Another advertising executive said: “What level of targeting that they start to apply could get really, really interesting. We were keen for some of our brands to use it so that we could see what was coming [but] we can’t correlate exposure to Netflix to sales — it’s just too small. But you can get learning about who saw your ads.”

Netflix is also seeking to build out its gaming business to include advertising, noted another ad executive.

Others added that they expected Apple to follow Netflix and introduce advertising alongside its shows and products.

Netflix and Microsoft declined to comment.

Initial hopes of 1mn shipments in 2024 launch year dashed by manufacturing problems

Apple has been forced to make drastic cuts to production forecasts for the mixed-reality Vision Pro headset, unveiled last month after seven years in development and hailed as its most significant product launch since the iPhone.

The complexity of the headset design and difficulties in production are behind the scaling back of targets, while plans for a more affordable version of the device have had to be pushed back, according to multiple people with direct knowledge of the manufacturing process.

Apple has already flagged that the $3,500 “spatial computing” headset device will not go on sale until “early next year”, a lengthy gap from its June 5 launch.

Analysts have interpreted this as being more to do with supply chain problems than allowing developers time to create apps for the Vision Pro.

Two people close to Apple and Luxshare, the Chinese contract manufacturer that will initially assemble the device, said it was preparing to make fewer than 400,000 units in 2024.

Multiple industry sources said Luxshare was currently Apple’s only assembler of the device.

Separately, two China-based sole suppliers of certain components for the Vision Pro said Apple was only asking them for enough for 130,000 to 150,000 units in the first year.

Both projections imply a significant cut to production from an earlier, internal sales target of 1mn units in the first 12 months. The forecasts for low volumes reflect Apple’s lack of confidence in being able to scale production, according to analysts and industry experts, following years of missed deadlines in launching the device.

Apple, whose market valuation closed above $3tn on Friday, three weeks after the announcement of the headset, declined to comment on the Vision Pro.

Luxshare did not respond to a request for comment.

Among the major hurdles being faced is the manufacturing of the sleek screens for the device. They consist of two micro-OLED displays — one per eye — and an outward-facing, curved “lenticular” lens. The inward displays offer a resolution exceeding anything currently on the market, while the outward lens projects the headset wearer’s eyes to the outside world.

The micro-OLED displays for the prototypes in the June demonstration were supplied by Sony and the chipmaker TSMC, according to two people familiar with the situation. Sony and TSMC declined to comment on any role in the Vision Pro.

Apple has been unhappy with suppliers’ productivity, said those people, especially with the yield of micro-OLEDs that are free of defects. The displays are the most expensive component in the Vision Pro.

“A lot of this is normal growing pains,” said Jay Goldberg, founder of tech consultancy D/D Advisors. “This is the most complex consumer device anyone has ever made.”

Goldberg said the higher than expected $3,500 price point already implied that Apple had baked in the cost of production inefficiencies, knowing that manufacturing yields were especially low compared with the mature products in Apple’s portfolio.

“Someone has to pay for that,” he added. “I think Apple went into this with a lot of ‘bad yield’ built into the model. There is a lot of technology in the Vision Pro and they knew it would take a while to scale up. Apple knows they won’t make money on this in the first year.”

Sony was cautious about how much the mixed-reality headset market would expand and was reluctant to step up production significantly, said Terushi Shimizu, head of Sony’s semiconductor unit, in a recent media roundtable.

“We will be watching to see how much demand [for micro-OLED displays] will increase,” he said. “But I don’t think we will be aggressive [in producing] in the same scale as image sensors”, for which Sony is building a new plant to boost production for the chips used in smartphone cameras.

Meanwhile, Apple is already working on later generations of the headset, including a more affordable version that is expected to appeal more to mass-market consumers, said two people with direct knowledge. 

Apple is working with Korean display makers Samsung and LG on this second-generation headset. In order to drive the price lower, the iPhone maker has explored using other display technologies, including mini-LED, but two people said Apple was insisting on using micro-OLED even for the non-Pro headset, even though all suppliers had so far failed to match its expectations.

The cut to 2024 forecasts has disappointed Luxshare, which has been gearing up its capacity to be able to build nearly 18mn units annually in the coming years, according to one person close to the assembler. 

Analysts said the whole headset supply chain in Asia was not getting much of a boost from the Vision Pro. “Apple has not made a better product than the industry imagined . . . the manufacturers’ confidence is not high,” said Eddie Han, an analyst at Taiwan-based Isaiah Research. 

Despite the setbacks, market intelligence group Canalys believed Apple would surpass a user base of 20mn within five years of launch.

“Given the limited production numbers, it will be flying off the shelves, pre-ordered by Apple’s loyal fans and high net worth users in the US,” said Canalys analyst Jason Low. His projection was currently for Apple to produce 350,000 units next year, increasing to 12.6mn units five years later.

It was a very busy week in the world of fintech, which certainly kept us on our toes. We covered a couple of notable M&A deals (including one of the biggest of the year so far), a different kind of financial services startup aimed at undocumented immigrants, Brex’s official recommitment to the startup community and more.

End-of-quarter frenzy

As the generative AI craze rages on, Ramp acquires customer support startup Cohere.io

We started the week with some M&A news out of corporate spend management company Ramp. The team shared with us exclusively that it had scooped up an AI-powered customer support called Cohere.io, which had raised $3.5 million in seed funding over its lifetime from backers such as Initialized Capital, Y Combinator and…Ramp co-founders Eric Glyman and Karim Atiyeh. Notably, Ramp (and companies like Deel and Rippling) were also customers. Glyman told us that it was evident from early on that once his company started using Cohere.io, “suddenly the majority of tickets were being answered properly in an automated fashion. […] It actually really worked,” he said. “The technical sophistication of the team was far beyond anything we had ever seen.”

Visa acquires Brazilian fintech startup Pismo in $1B blockbuster deal

Then later in the week came confirmation of one of the biggest, if not the biggest, fintech M&A deals of the year.

It was rumored for months that Visa, Mastercard and potentially a bank and private equity firm were all courting Brazilian payments infrastructure startup Pismo. The acquisition was definitely a coup for the Latin American startup community, considering that Visa could have likely considered companies from all over the world. Pismo has apparently seen some explosive growth in recent years — jumping from 10 million accounts at the end of 2020 to 80 million today. Also, at the beginning of 2021, Pismo was doing less than $1 billion per month in transaction volume compared to processing $40 billion in transaction volumes annually today.

However, as noted by KBW managing director Sanjay Sakhrani, the $1 billion purchase price is roughly 30% below the $1.4 billion that Visa was rumored to have offered for Pismo earlier this year.

We don’t know what Pismo was valued at when it raised $108 million in a round co-led by SoftBank, Amazon and Accel in 2021.

But Accel partner Ethan Choi told us the sales price was “a very strategic multiple.”

Sakhrani also said in a report that in addition to beefing up Visa’s issuer processing capabilities across card products, Pismo also brings “differentiated core banking capabilities and will allow Visa to provide connectivity and support to emerging payment rails like Pix in Brazil.”

The last week in recent memory where we remember seeing such a flurry of fintech M&A activity was in mid-January, when Jonah Crane, partner at Klaros Group, predicted we would continue to see more acquisitions in 2023 thanks to the continued venture slowdown and practically dead IPO and SPAC markets. And according to CB Insights, fintech M&A exits rebounded in the first quarter, but not as much as one might have expected. They were up 15% QoQ to 172 deals. Most of Q1’23’s top M&A deals involved fintechs based outside of the U.S. For the first time in the previous year, the top M&A valuation fell below $500 million.

Side note: The acquisition represented a rare win for SoftBank, which has had a number of high-profile disappointments in recent years with investments in the likes of WeWork, the now defunct Katerra and FTX. Alex and I talk more about that on Friday’s episode of Equity Podcast here.

Brex refocuses on startups with hire of SVB veteran, ex-a16z operating partner

Last summer, Brex made headlines for announcing it would stop serving SMBs and non-funded startups. This summer, it’s making headlines for pledging its recommitment to the startup community. After Silicon Valley Bank imploded in March, Brex (along with the likes of Arc and Mercury) saw an influx of new customers.

Specifically, the company says it opened 4,000 new accounts and received $2 billion in deposits in the first week after the SVB shutdown alone.

That obviously led the company to rethink its strategy. Last week, Brex told us exclusively that it had hired Jason Mok, a former operating partner at Andreessen Horowitz (a16z) and 16-plus-year veteran of Silicon Valley Bank to serve as its head of startups. I talked with Mok about his previous experience and how he thinks that will help him in his new role, which includes providing more “Brex ambassadors” who can serve as the face of the brand that founders, operators and VCs can go to for advice, perspective and connections to other founders.

No SSN, no problem, says Maza, a fintech startup aimed at undocumented immigrants

I also wrote about Maza, a fintech startup that raised $8 million in a seed-funding round led by a16z to help undocumented immigrants get an ITIN (individual tax identification number) and access the U.S. financial system. TechCrunch has previously reported on a number of startups focused on the immigrant community — including Welcome Technologies, Fair, Majority and TomoCredit. (It’s unclear whether Fair is still around considering its website appears to no longer exist or is down.) But what makes Maza different is its focus on undocumented immigrants specifically. “We’re doing much more than just providing a bank account — we’re really giving immigrants a stable and legal financial foundation from which to build credit and wealth indefinitely,” said co-founder and COO Robbie Figueroa.

Fun fact: The name Maza came from a song about perseverance called “La Maza” that co-founder and CEO Luciano Arango used to listen to with his mom growing up.

Nubank’s CEO explains what the US could learn from LatAm fintech

In one of the Equity podcasts this week, I spoke with David Vélez, the co-founder and CEO of digital bank giant Nubank.

(Did you know it has a market cap of $37 billion?!)

Digital banking is always top-of-mind over here at TechCrunch, so we quickly got into discussing Nubank’s strategy for achieving profitability and how the company has been able to maintain that in a challenging macroeconomic environment.

Vélez also compared and contrasted the Latin American and U.S. fintech markets and dished on how he sees banking evolving in the next few years. One particularly interesting part of the conversation, as one Equity listener pointed out: Vélez’s reasoning for only expanding Nubank in three markets over 10 years: “‘I am extremely wary of any deck where people tell me ‘18 markets in 2 years.’”

Fidelity Investments valued its stakes in Discord and Reddit slightly lower than it had previously, according to new public disclosures reported by TechCrunch.

The asset manager marked down the value of its Discord holdings by 13% at the end of May, compared to the previous month. At the same time, it marked down its Reddit stake by about 7%. Those declines come after larger markdowns last year amid a steeper slump in public tech stocks.

Twitter, on the other hand, got better news. Fidelity marked up its stake by about 5% in May compared to April.

THE TAKEAWAY

• OmniML helps complex models run on devices

• The deal could help Nvidia improve its chip for autonomous vehicles and robots

• Nvidia faces competition from smaller chipmaker in edge computing

Nvidia in February quietly acquired OmniML, a two-year-old artificial intelligence startup whose software helped shrink machine-learning models so they could run on devices rather than in the cloud, according to a spokesperson and LinkedIn profiles of former OmniML employees who now work at Nvidia.

The acquisition could be a sign that the chipmaker, whose data-center server chips have fueled a recent AI boom and enabled chatbots including ChatGPT, wants to improve its separate AI chips for cars, industrial robots and drones. The startup’s engineers also could potentially aid an effort to shrink the AI software that powers chatbots so it can run on devices rather than in data centers.

Terms of the deal weren’t disclosed. Since the acquisition, Nvidia has become an increasingly active investor in hot AI startups that buy its chips. In recent weeks, for instance, it bought equity stakes in chatbot developer Inflection AI and video-editing software firm RunwayML. (See our Generative AI Database.)

Nvidia designs most of the world’s graphics processing units, which have been particularly useful for developing AI software such as large-language models, the technology behind ChatGPT.

Now developers of LLMs, including ChatGPT creator OpenAI, are increasingly working on ways to make the software smaller so it can run on industrial or personal devices such as smartphones or laptops. Running such software locally on a device is typically faster and less expensive than running it in the cloud.

It isn’t clear whether the OmniML technology or personnel would aid in the effort to shrink LLMs. But the LinkedIn page of one of the startup’s founding engineers says he now works at Nvidia on “model and pipeline optimization for large generative models.”

The startup previously said its software could compress the size of machine-learning models so they could power AI on devices, but the examples it cited primarily involved computer vision—helping a smart camera or autonomous vehicle identify objects around it, for example.

The OmniML technology would fit in with Nvidia’s existing business of selling chips that process data on vehicles and industrial devices so they can handle such tasks. Nvidia does not break out financials around that business, known as edge computing.

Nvidia has cornered the market on GPUs for data center servers, but it faces some competition in edge computing chips.

Startup SiMa.ai, for instance, is working on AI chips for devices ranging from robots to cars to cameras. The startup has raised $200 million in venture capital and recently beat Nvidia in a closely watched test, called MLPerf, of chip power and performance.

Nvidia declined to comment on when the OmniML purchase closed and how much it paid to buy the startup. OmniML said it raised $10 million last year in a seed round led by GGV Capital that also included Qualcomm Ventures and Foothill Ventures. In January, OmniML announced a strategic partnership with Intel, an Nvidia rival, to bring its software to Intel hardware for some customers.

That partnership may have ended following the Nvidia purchase.

Intel did not immediately respond to a request for comment.

OmniML was founded by Song Han, an electrical engineering and computer science professor at the Massachusetts Institute of Technology; Di Wu, a former software engineer at Meta Platforms; and Huizi Mao, who co-invented “deep compression” technology, which came out of Stanford University and is at the heart of the startup’s software.

Wu and Mao, who did not immediately respond to a request for comment, now work at Nvidia, according to their LinkedIn profiles.

“AI is so big today that edge devices aren’t equipped to handle its computational power,” Wu said in a March 2022 press release.

Nvidia plans to use OmniML’s technology to help Nvidia’s customers develop AI models faster, including by improving accuracy and reducing latency for complex machine-learning models, according to the spokesperson.

The acquisition happened three months before Nvidia stock jumped 25% following its announcement of a blowout sales forecast for the current fiscal quarter, which ends in July, owing to exploding interest in LLMs.

Nvidia has a market capitalization of $1 trillion.

Regime stability is a funny thing. One day it’s there; the next day, poof—it’s gone.

For the past several months, I have been talking to experts about a possible coup in Russia. I approached the question gingerly. It seemed too much to hope for; it seemed naïve. Vladimir Putin had been in power for more than two decades. Many had predicted his demise—always prematurely. There was a small cottage industry on Twitter of people insisting that Putin was ill. They liked to post photos of him sitting at meetings, clutching his desk as if he were about to fall. I didn’t want to be like that. “Is this ridiculous to even think about?” I would ask the experts. The experts laughed. They felt the same way. A coup was unlikely, they agreed. A popular uprising—a “Ceaușescu scenario,” in which the people stormed the Party’s headquarters, convened a hasty trial, and murdered their dictator—probably even less so. To a scenario like the one that actually played out last weekend—one of Putin’s warlords raising a mutiny, taking over one of the country's military headquarters, and marching on Moscow, all while Putin was still in power—we gave very little consideration. It just seemed too outlandish to talk about.

And yet, since the war began, all of the experts had been thinking about ways in which the Putin regime might collapse, and watching what Putin was doing to protect himself. Peter Clement, a former director of Russia analysis at the C.I.A., noted a televised meeting, days before the war, in which Putin browbeat members of his security council into pledging their support for his Ukraine policy. It was a brilliant move by Putin, Clement thought, to bring his senior administration officials in line. “They’re all complicit now,” Clement said. “It’s not like one of them can say, ‘I thought this was a stupid idea.’ They all signed on.”

For that reason, Clement thought it more likely that a move against Putin would come from the second circle, from someone less in the public eye, someone we’d not heard of. Clement was willing to speculate with me, but he considered the chances low. You’d have to have the security services on board, he said, because you’d need to physically arrest the President, and it was unlikely you could appeal to security hawks with an antiwar agenda. And you’d have to be prepared to run the country. It’s a big country and in the thick of a long war. “It can’t just be, ‘We got rid of the Wicked Witch of the West! Let’s all stand up and cheer!’,” Clement said. You’d have to have a plan, and Clement was having trouble thinking of people who might have one.

Another former C.I.A. analyst, Andrea Kendall-Taylor, who was a deputy national intelligence officer for Russia and Eurasia between 2015 and 2018 and now runs the Transatlantic Security Program at the Center for a New American Security think tank, walked me through the political-science literature on how authoritarian regimes tend to fall. Of the four hundred and seventy-three authoritarian regimes that had fallen between 1950 and 2012, a hundred and fifty-three had done so via coup. But the coup was on the wane; after the end of the Cold War, the U.S. had stopped propping up quite so many military dictatorships, which are what tend to get militarily couped. It was unlikely, Kendall-Taylor explained, that the security services, or anyone from Putin’s inner circle, would move against the Russian President, because the regime had entered the stage that the political scientist Milan W. Svolik called “established autocracy.” In an established autocracy, the leader has monopolized power to such an extent that he can no longer be threatened by what Svolik calls an “allies’ rebellion.” The truth is, Kendall-Taylor said, most personalist dictatorships, such as Putin’s, ended with the dictator dying in power, especially when the dictator was older than sixty-five (Putin is seventy). “That is by far the most likely scenario,” she told me. She put the chance of regime change in Russia in the next two years at ten per cent, and that “ten per cent includes Putin having a heart attack.”

The historian Vladislav Zubok, who is the author of a recent book on the collapse of the Soviet Union, described the various ways in which other Russian and Soviet leaders—Nicholas II, Nikita Khrushchev, Mikhail Gorbachev—had been ousted, and explained why none of those scenarios mapped onto this one.

Nicholas II had abdicated, in 1917, after large protests in Petrograd (current-day St. Petersburg, then the Russian capital) shattered confidence in his regime, and the military joined the mutiny; Putin, Zubok pointed out, had made sure that his capital, Moscow, was well-provisioned and maximally isolated from the war in Ukraine; there is a loyal paramilitary force to control protests.

Khrushchev was overthrown, in 1964, by a plot from within his own inner circle, led by his deputy, Leonid Brezhnev, who worked within the structures of the Communist Party to urge others to turn against their leader.

The K.G.B. played a key role in the coup. Putin’s regime, by contrast, is highly informal, much more like Stalin’s, with all paths leading, in the end, to Putin. It is hard, under such circumstances, to plan a coup. And there are several branches of secret police, each competing with the others, making any plotting very complicated. As for Gorbachev, the comparison seemed the least apt of all. He had not only allowed his rival, Boris Yeltsin, to run for President of Russia—he permitted the government to finance his campaign.

Putin was unlikely to do something like that.

If there was a leader to whom Putin could be compared, Zubok said, it was Ivan the Terrible, who ruled Russia in the second half of the sixteenth century. Ivan fought a long war of attrition with his Western neighbors; he demoralized his ruling élite, and murdered his own son and heir. After his reign was over, the country eventually fell into civil war, the period known in Russian history as the Smuta, the Time of Troubles.

Two experts on Russian public opinion described their understanding of Russian attitudes toward the war in Ukraine, and what might cause those attitudes to change. Oleg Zhuravlev, a founding member of the Public Sociology Laboratory, an independent Russian research collective, summarized a series of in-depth interviews that his team had done with young Russians in the past year. They had found that support for the war was both thinner and narrower than it looked. There was a small group, about ten to fifteen per cent, of genuine supporters; there was a similarly small group of genuine opponents. In between was a large group of people, most of whom had come around to supporting the war not because they thought it was a good idea but because they didn’t know how to oppose it, and because they felt totally alienated from the people in charge of it. “Over and over we heard the same thing,” Zhuravlev said. “ ‘If there’s one thing I know about politics, it’s that I don’t know anything about politics. The people in the Kremlin are foreign to me; they are not like me. But they must have their reasons.’ ”

It was depoliticization in its purest form. Zhuravlev’s occasional collaborator, the longtime polling expert Elena Koneva, had spent the year and a half since the war began running a project called ExtremeScan, through which she designed polls to figure out the basis for Russian public support of the war and what could cause it to contract. She had seen signs, mostly in the border regions of Russia, that, when the war began to truly affect people’s lives, their opinions started to change. First they experienced fear of retribution—“We have done so many horrible things to Ukraine,” one respondent said, “that the Ukrainian Army will inevitably come here”—but the actual experience of war, of shortages, of shelling, of people being forced to evacuate, began to erode support for the war. And Koneva predicted that, if things got worse, support would erode further. “If people are constantly having to sit in bomb shelters, and women are giving birth without medicine,” she said, “then an end to the war will become their most passionate wish.”

Yevgeny Prigozhin figured in our conversations as a grotesque and somewhat comic character. When looking at the Putin regime, one Moscow-based historian said, “We’re all wondering who the Beria figure is going to be,” referring to one of Stalin’s most efficient henchmen, tried and executed by his former comrades after Stalin’s death. “Who are they going to take out and shoot right away? And then you look at the criminal types who are working for the Kremlin—and you see Prigozhin. There’s your Beria.”

For Kendall-Taylor, speaking in May, Prigozhin’s antics—his profane insults and increasingly aggressive rants, which included accusations of treason against the Russian Army’s leadership—were a sign of élite discord. In a post-Putin world, she said, the presence of warlords like Prigozhin and Ramzan Kadyrov, the head of the Chechen Republic, could lead to a “Sudan scenario,” in which these forces would start a civil war. In the near term, though, with Putin still in power, she did not think Prigozhin would undertake an actual rebellion. At the time, his criticisms of the military seemed only symbolically significant, a sign that the élite was in disarray and that protest actions, whether secessionist or antiwar, might not be met with as much force as people had once thought.

Regime stability is a funny thing.

One day it’s there; the next day, poof—it’s gone.

The Moscow-based historian, who asked that his name not be used since he was still in Russia, recalled what it was like to observe the Politburo in the early nineteen-eighties. “They looked like a totally homogeneous mass,” he said. “There was no indication, in their public statements or in anything else, that any of these people thought differently from one another.” But Gorbachev, it turned out, did think differently. In the years to come, he undertook a series of reforms that ended with the Soviet Union ceasing to exist.

Authoritarian regimes could seem very stable, until suddenly they weren’t.

On the first anniversary of the Russian invasion of Ukraine, Kendall-Taylor convened a group of experts to compile a “stability tracker” for the Putin regime. The tracker identifies ten “pillars,” ranging from “Absence of an alternative to Putin” to the idea, among Russian citizens, of “Russia as a besieged fortress,” and tries to indicate whether these are growing stronger or weaker. As of this spring, several factors were going in the wrong direction for Putin: his élite was becoming fragmented; his economy was suffering the effects of the war and of sanctions; and his military, historically apolitical, was being pulled into the political arena by concerns over Prigozhin’s rising influence and its access to military resources. But the factors going in the other direction were more numerous: according to Kendall-Taylor’s experts, Putin had strengthened his control over the information environment; the people most discontented with his rule were leaving the country; and the idea of Russia as a besieged fortress was gaining rather than losing adherents. Most important, there remained no viable alternative to Putin: his warlords were politically unpopular, and his heroic opponent, Alexey Navalny, was being denied food, sleep, and medical care in a Russian prison.

In the absence of an alternative, the status quo would continue.

Among experts thinking about the Russian regime, there are, very roughly speaking, two kinds: those who look at Russian and Soviet history and culture to determine what might happen next, and those who look at Russian authoritarianism in comparative perspective—that is, alongside authoritarian regimes in Egypt and China and Turkey. This is also known as the political-science-versus-area-studies debate. The different approaches yield slightly different hypotheses. Before joining the C.I.A., Clement wrote a Ph.D. dissertation on the so-called Congress of Victors, in 1934, at which Stalin consolidated his rule and also began to see that not everyone was satisfied with it. In the eighties, Clement analyzed the succession struggles during the post-Brezhnev period. His interest in regime insiders partly stems from this experience. Kendall-Taylor, a comparativist with a Russian focus, who studied in graduate school with Barbara Geddes, one of the founders of modern quantitative authoritarian studies, prefers looking at the numbers: this many regimes of this type fell in this manner; this many regimes of this different type fell differently. But everyone agrees that the life of a regime is full of contingencies; leaders can make mistakes. They can, for example, start a brutal and senseless war against a neighboring country, and refuse to relent even when the war is going badly.

War is a known stressor for personalist dictatorships, which (paradoxically, one might argue) also tend to start more wars. War puts pressure on the economy and on security services, and also has a way of being unpredictable. Between 1919 and 2003, according to the political scientists Giacomo Chiozza and H. E. Goemans, nearly half of all rulers who lost wars then lost power within a year. (Of these, half were sent into exile and nearly a third were jailed.) This is true of Russian history as well. War losses, such as Russia’s 1905 defeat by Japan, have sometimes led to tectonic shifts in the country’s political life; in 1905, it led to an uprising that forced the tsar to grant his people a constitution. In 1917, the Russian Army’s struggles in the First World War were a major factor in pushing the tsar out of power.

But wars have an upside for authoritarians, too. They offer an excuse to increase repressions and take over the information space. And repression, unfortunately, works. It makes people more fearful of voicing their dissent and of coming out into the street. In the current situation, Kendall-Taylor gave credit to the work of the Russian security services, who have tended to respond to protest with some restraint. “They don’t overreact in ways that could spiral and trigger a public reaction,” she said.

“They’re using facial recognition and other things.

It’s a knock on the door out of the public eye rather than beating people in the square.” This tends to mitigate some of the dangers of public disapproval that are inherent to a repressive system.

On the other hand, repression’s ability to decrease the flow of information can be dangerous for a regime. People don’t know how dissatisfied other people are—and neither does the government. Kendall-Taylor recalled Mohamed Bouazizi, the Tunisian fruit vendor who set himself on fire in December 2010 and started the Arab Spring. This was an example of the downside of repression. Repression works, until it doesn’t. In Tunisia, one desperate act led to a popular uprising and the fall, in weeks, of a regime that had been in power for decades.

Prigozhin’s march to Moscow came as a shock to just about everyone. Clement had been on high alert ever since the Russian authorities declared, on June 10th, that they would require all Wagner soldiers to sign a contract with the Ministry of Defense—that is, technically, to dissolve Wagner as an independent entity. Zubok had also noted the increasing fractiousness of the Russian élite, and wrote to me before the uprising that the emperor might not be clothed. Kendall-Taylor, in an article she co-wrote for Foreign Affairs, mentioned Prigozhin as a possible pretender to the throne after Putin’s departure. But no one expected the sequence of events that unfolded last weekend.

In their aftermath, there were more questions than answers. Zubok wrote a piece for the New Statesman in which he compared Prigozhin’s act to Julius Caesar crossing the Rubicon and marching on Rome in 49 B.C.; he recalled that the oligarch Mikhail Khodorkovsky had once proclaimed, when he was a young man on the make in the new Russia, that the man with the rifle—an iconic symbol of the Bolshevik Revolution—had been replaced by the man with the ruble. But now the man with the rifle was back; a new political era had dawned in Russia.

I asked Zubok why he had decided to go outside of Russian history to try to explain the Prigozhin phenomenon. He said that Russian history can sometimes feel like a straitjacket. “ ‘Hey, it’s the Smuta! Hey, it’s 1917! Hey, it’s 1991.’ I’m the last person to argue that those patterns don’t matter.” But sometimes you want a little something different. Sometimes you want to explore other historical connections.

But, he added, “there is a huge caveat.” Despite its many wars and campaigns, Russia has never actually produced a Caesar—that is, a warlord who marches on the capital with his men and takes political power. There is a reason for that. The traditional political system, arguably still in place to this day, is a triangle comprising the tsar, the boyars, and the people. In times of trouble, the tsar can play the people against the boyars, and vice versa. If things go bad, the boyars can take the blame. This is, in effect, what Prigozhin was asking for—that Putin sack his boyars in the Army, who had made such a hash of the war. Zubok acknowledged the danger inherent in such a strategy for Putin: “You may think that someone is a Red general, but next thing you know they’ve turned around and are executing the Bolshevik leadership”—as happened with Ivan Sorokin, a revolutionary commander in the Russian Civil War who went rogue in the North Caucasus and attacked the Soviet leadership in his own district before finally being killed himself. But this is the sort of thing that happens in the absence of a tsar, when the Smuta is in full swing, whereas Putin, however weakened, remains the tsar. “You have to acknowledge the sources of resilience in this crazy system,” Zubok said. His prediction was that Putin would remain in power, chastened but basically unchanged.

This was also Clement’s analysis. There were many things about the events of the weekend that he found notable, including the fact that Prigozhin’s column of trucks and armor, travelling, exposed, along Russia’s highways, had managed to go as far as they did. “When was the last time you saw a column on a highway?” Clement asked, recalling the Russian trucks and tanks on the road to Kyiv last year that had been methodically picked apart by Ukrainian forces. His conclusion was that local commanders did not feel that they could take the initiative to destroy Prigozhin’s column; it was above their pay grade. He was also fascinated by Putin’s five-minute video address, recorded during the uprising, in which Putin spoke of treason and betrayal and seemed to compare himself to Nicholas II, unable to prosecute a war because of intrigues behind his back. “Was it absolutely necessary to make this speech?” Clement asked. It made Putin look panicky and weak, he said: “If you really think it’s a full-blown rebellion, why don’t you take him out?”

Nonetheless, he could see no pathway to a Russia without Putin. An analysis in the Times had suggested that there could be talk in his inner circle of asking Putin not to stand for reëlection in 2024. Clement was skeptical. “The trouble with that is, who’s the person who’s going to go in there and say that to him? Who is going to say, ‘Vladimir Vladimirovich, look: you are a very rich man—we think you should go and retire and just live happily ever after’?” Clement recalled an incident during the Iran-Iraq War, in which one of Saddam Hussein’s advisers suggested that a way to forge peace would be for Saddam to temporarily step down as Prime Minister. The man was executed and his body parts delivered to his family the next day. “Dictators don’t like to be told that they should retire,” Clement said. He added that, in this, they weren’t necessarily wrong: Could Putin actually retire? Who could guarantee his safety? Wouldn’t whoever replaced him as the ruler of Russia find it very uncomfortable to have Putin still hanging around? “This isn’t like Khrushchev, where he can just go live quietly on his farm,” Clement said. This was a person with a lot of enemies.

For the moment, at least, it was also a person with a lot of power. “He still controls the Army and the F.S.B.,” Clement said. “And people are still afraid of him.” Clement suspected that Prigozhin would meet an unhappy end. Putin is notably vengeful, widely believed to have approved the murder of people he deemed to have betrayed Russia, long after they had done it, though the Kremlin has repeatedly denied the country’s involvement. But Clement also believed that the experience with Prigozhin could make Putin more cautious: he might make some changes among his advisers, as a way of explaining the failures of the war; he might even conclude, Clement speculated, that the war was taking away too many resources, that he needed to focus on domestic concerns, and that he would therefore consider engaging in ceasefire talks, so that he could regroup and possibly resume war later. That, Clement went on, was still low probability. But the probability had increased.

Zhuravlev, the sociologist, observed Prigozhin’s uprising from a hospital in Almaty, Kazakhstan, where he had gone to do research and had come down with appendicitis. He found the spectacle fascinating—both terrifying and encouraging—but the videos of ordinary Russians in Rostov-on-Don greeting Prigozhin’s fighters and cheering them on did not entirely surprise him. Zhuravlev interpreted it, optimistically, as a sign of engagement. In his interviews with young Russians, he had been struck by how many of them wanted to talk about the war but had no opportunities to do so. Now there was an opportunity—not owing to a democratic movement, to be sure, but an opportunity nonetheless. The uprising was encouraging in another way, too: it showed that it might not be so hard to organize a revolution in Russia. “For now, the people who want to do it don’t have the means,” Zhuravlev said, “and the people who have the means don’t want to do it.” But perhaps this could change.

Kendall-Taylor was quick to clarify that the events of the past weekend were not an attempted coup but an insurgency, a sign of frustration rather than a planned attempt at regime change. Nonetheless, the message that Prigozhin’s actions had been sending for months—that the regime was not as strong as it seemed; that you could defy it and survive—had been significantly strengthened. “So much of the glue of these regimes is that no one knows how widely held the discontent is,” she said. “When something like this happens, it sends such an informative signal that others are as dissatisfied as you are. It starts to change people’s calculus about what is possible.” The next time something happens in Russia that people do not like—it could be a major military defeat, or, slightly more likely, according to Kendall-Taylor, something to do with the 2024 Presidential election—they may not be so worried about going out into the street to say so. Prigozhin may or may not survive his stay in Belarus, but here was a person who marched with a small army several hundred miles through Russia, without encountering any real resistance. That didn’t mean the regime was in danger of imminent collapse—but it suggested that the chances had increased slightly. “That’s the way these regimes unravel,” Kendall-Taylor said. “At the end of the day, whether it comes from a coup or an insurgency or a protest, Putin will at some point give an order to crack down and fire, and people won’t do it. And that’s the end of the regime.”

No one could predict the future. But it was worth trying to analyze the situation and think it through. Earlier, Kendall-Taylor had said that the chances of Putin no longer being in power in two years were ten per cent. Now she was willing to go up to twenty.