‘Yellowstone’ Creator Taylor Sheridan to Leave Paramount for NBCUniversal
Sheridan’s hits for Paramount include ‘Tulsa King’ and ‘Landman.’ In addition to TV shows, he will make movies for NBCUniversal.
Taylor Sheridan will join NBCUniversal after his Paramount contract expires.
Sheridan will create films for Universal Pictures and TV content for NBC and the Peacock streaming service.
NBCUniversal also signed a first-look deal with Sheridan’s production partner 101 Studios, effective in 2026.
Prolific movie and television producer Taylor Sheridan, known for hits including “Yellowstone,” “Tulsa King” and “Landman,” has signed a deal to join NBCUniversal when his contract with Paramount expires, according to people familiar with the matter.
Sheridan has become one of the most reliable creators and producers for Paramount over the past several years and has been one of the biggest providers of shows for the Paramount+ streaming service.
Sheridan’s departure is a potentially big blow to Paramount. However, he will stay involved in all his shows there and won’t depart for NBCUniversal until 2029, after his current contract expires, the people familiar with the matter said.
At NBCUniversal, Sheridan will create films for its movie studio Universal Pictures as well as TV content for NBC and the Peacock streaming service.
News of Sheridan’s eventual exit was earlier reported by Puck.
Sheridan’s decision to leave comes less than three months after David Ellison’s Skydance Media acquired control of Paramount from Shari Redstone’s National Amusements.
Ellison previously said in an interview with CNBC soon after closing the Paramount acquisition that he hoped to keep Sheridan making shows for the company for many years to come.
Known for shows featuring old-school and frequently older male heroes often confronting changing times, Sheridan was initially seen as making shows that appealed more to middle America than coastal elites. However as his shows have become more popular, his audience in urban markets grew.
Besides his television shows, Sheridan’s film credits include “Sicario” and “Hell or High Water.”
While his shows are successful in the ratings, they are also costly to produce. In 2023, a Wall Street Journal article detailed the inner workings of Sheridan’s operation including how the writer-producer has far more freedom than most producers. He often shoots at ranches he owns in Texas where he can bill Paramount as much as $50,000 a week, the Journal reported.
NBCUniversal has also signed a first-look deal with Sheridan’s longtime production partner 101 Studios that takes effect in 2026.
Two U.S. Navy Aircraft From Same Carrier Crash Into South China Sea
Incidents coincide with President Trump’s visit to Asia and are under investigation
A U.S. military helicopter and a jet fighter from the USS Nimitz crashed into the South China Sea within 30 minutes on Sunday.
All five crew members from both aircraft were rescued and are in stable condition; the cause of the crashes is under investigation.
The incidents occurred during President Trump’s visit to Asia and after the USS Nimitz’s deployment to the Middle East.
KUALA LUMPUR, Malaysia—A U.S. military helicopter and a jet fighter from the same aircraft carrier crashed into the South China Sea within 30 minutes of each other on Sunday.
The two aircraft’s five crew members were rescued and are in stable condition, the U.S. Pacific Fleet said on X. Both aircraft had taken off from the USS Nimitz, America’s oldest aircraft carrier that is returning to its home base on the U.S. West Coast for decommissioning scheduled for next year.
The cause of both crashes is under investigation, the fleet said.
The USS Nimitz had been deployed in the Middle East as part of the Pentagon’s response to attacks by Yemen-based Houthi rebels on commercial ships in the Red Sea.
The crashes coincide with a visit by President Trump to Asia and follow a series of mishaps over the past year on another aircraft carrier, the USS Harry S. Truman, during its deployment to the Middle East.
The Pacific Fleet said an MH-60R Sea Hawk helicopter went down in the South China Sea around 2:45 p.m. local time on Sunday while conducting what it said were routine operations. Thirty minutes later, an F/A-18F Super Hornet fighter crashed into the same waters, the fleet said. The jet’s two crew members ejected before the aircraft went down and, like the helicopter crew, were safely recovered by the carrier’s search and rescue teams.
The South China Sea is one of the world’s most contested waterways, where Beijing’s claims for territorial waters clash with those of several Southeast Asian nations. The U.S. military has worked with allies in the region to preserve freedom of navigation in the South China Sea, which is a thoroughfare for about one-third of global maritime trade.
Tuition fee rises may not stop universities in England going under, says regulator
Chair of the Office for Students told the FT most institutions were ‘on a journey to balancing income with expenditure’
A high-profile university could still collapse despite a rise in tuition fees “stabilising” the finances of higher education in England, the chair of the sector’s regulator has warned.
Edward Peck, chair of the Office for Students, told the Financial Times most universities were “on a journey to balancing income with expenditure” but that some institutions could yet fail, and called on vice-chancellors to strengthen governance to avoid insolvencies.
Peck’s comments come after the UK government last week laid out a long-awaited strategy for higher education, vowing to overhaul research funding, strengthen regulatory oversight and legislate for inflation-linked increases in tuition fees in a bid to improve the sector’s finances.
“Increasing fees linked to inflation has stabilised most institutions in their current position . . . but there is still risk,” Peck said, describing a cash crunch caused by alleged mismanagement at Dundee university this year as a “wake-up call” for the sector.
“The OfS has paid too little attention to governance,” he added, while urging universities to see themselves as “autonomous organisations” and improve financial oversight in order to minimise intervention by the regulator.
“That’s got to be our approach in the future, to be as collaborative as possible . . . There’s no benefit to them or us in having a stand-off . . . they need us more than we need them at the moment.”
Many of the proposed reforms in the post-16 education and skills white paper — which reiterated Prime Minister Sir Keir Starmer’s pledge that two-thirds of people enrol in a university or technical degree or an apprenticeship by the age of 25 — have been welcomed by universities.
But institutions have warned that plans to fund maintenance grants for domestic students through a levy on international student fees will dent the government’s drive to boost the economy and further damage their finances, already hit by the rise in employer national insurance contributions.
Peck said the white paper’s emphasis on “a shift back to vocational technical education” presented an opportunity to “rethink what it means to go to university”, with incentives for universities to specialise and align with the government’s industrial strategy.
Asked about the planned levy on international student fees, to be set out in the Budget next month, he claimed it would demonstrate the benefits of overseas students to the public but acknowledged it would require universities to make “difficult judgments” about raising international fees.
Overseas undergraduates are charged up to £38,000 a year in tuition, far above the £9,535 paid by their English peers.
The OfS, set up in 2018, has repeatedly criticised universities’ “overly ambitious” projections for international recruitment after the previous Conservative government tightened visa policy in a push to slash immigration, leaving some institutions on the brink of collapse.
Peck, a former vice-chancellor of Nottingham Trent university who joined the regulator in July, said that while the sector’s forecasts now showed more “realism” and worst-case scenarios had not come to pass, some providers would have to take further steps to ensure a “viable future”.
These would include new models, such as the proposed merger between the universities of Kent and Greenwich, Peck said, adding that he was aware of no other tie-ups at present but expected to see more in future.
Stricter governance regimes — a key concern of the white paper, which called for “diverse skills and capability” to reduce the risk of boards signing off “unachievable plans” — were also needed, he added.
“Some institutions have not got ahead of the problem fast enough, and persuaded themselves that their student projections or financial forecasts were going to be realistic when they weren’t,” said Peck, adding that the sector was quickly learning that “the professionalism needed to run a university is not that dissimilar to a large business”.
Although the OfS preferred a university-led approach, ministers and the regulator would “need to give more consideration” to shaping the sector if providers were slow to act on the white paper’s recommendations, Peck noted.
The OfS plans to take a more hands-on approach in some areas, such as franchising and exploitation of students by recruitment agents, and the white paper pledged to increase the regulator’s powers in law.
Peck said a stronger remit would allow the watchdog to root out “bad actors in the system” more quickly.
Delivering the government’s ambitious reforms would probably require a bigger budget, but the OfS had set efficiency targets that would allow funding to be recycled into frontline regulation, he added.
Fees from regulated institutions accounted for most of the body’s £33.2mn budget in the 2024-25 financial year, with an extra £1.5mn in government funding to support its work on financial sustainability.
The OfS has been criticised for its pro-competition philosophy and lack of intervention, with a highly critical review last year calling on it to take a “more proactive” approach to “emerging risk”.
Peck admitted that there was “a lot of truth” in the idea that the OfS had “misunderstood distance for independence” and vowed to better engage with the sector, noting that it did “a great job for students most of the time”.
Chinese drugmakers shift from copycats to global innovators
Reforms, faster development and lower costs have turbocharged growth
Chinese drugmakers have struck a record number of deals this year to sell their assets overseas, sending biotech stocks surging as investors bet the country will become a driving force in global drug discovery.
There have been 93 overseas licensing deals in the first eight months of 2025, worth a combined $85bn, according to Chinese data provider PharmCube — with domestic companies selling the international rights to homegrown drugs. The total value of deals is on track to hit its sixth consecutive annual record, underscoring the speed at which the country’s pharmaceutical sector is internationalising.
Meanwhile, the rapid growth of clinical trials and drugs in early-stage development from companies, including Jiangsu Hengrui and Akeso, indicates that the trend of Chinese biotechs selling their drugs to global pharmaceutical companies is set to continue — a stark reversal from a decade ago.
“Ten years ago, China didn’t have a biotech sector to speak of. For the most part, the companies were developing generic drugs. Fast forward to today, and every big pharma is doing most of their shopping in China for novel therapies,” said Brad Loncar, an expert on Chinese biotech.
Beijing introduced reforms in the mid-2010s that made it easier for biotech companies to raise capital and pursue innovation. These changes, coupled with the relative speed and low cost of doing drug development and clinical trials, have turbocharged the industry’s growth. Hong Kong’s Hang Seng biotech index has surged more than 80 per cent this year on growing investor excitement.
Jiangsu Hengrui
China’s transformation from a copycat manufacturer of drugs developed overseas to a hub of homegrown research is exemplified by Jiangsu Hengrui. Founded in 1970, it spent the first two decades as a small-scale state-owned manufacturer of low-cost antiseptics. In the 1990s, it started developing generic anti-cancer drugs. It was privatised in 1997 and began investing in building its own research capabilities. Today, it has one of the most diversified pipelines in the country, spanning weight-loss therapies, oncology drugs and Alzheimer’s treatments.
Tony Ren, Head of Asia Healthcare Research at Macquarie Capital, said Hengrui was “a bellwether” for China’s pharmaceutical industry.
“The company offers a rare combination of proven R&D processes, a number of global outlicensing deals, China-centric sales skills, professional management and sound corporate governance,” he said.
Hengrui has struck licensing agreements with Merck, Braveheart Bio and Glenmark in the past year alone. In July, it agreed a deal with UK pharmaceutical company GSK to develop up to 12 medicines.
“In China, you can scale and test medicines in human beings much faster than in the US or Europe,” said Loncar. “If you have an idea for a drug, you can get an answer about whether it works a year or two earlier in China.”
For many Chinese biotechs, the surge in international partnerships has provided much-needed capital after a difficult few years marked by drug pricing reforms that squeezed profit margins on domestic sales.
Zhao Bing, a pharmaceutical analyst at Huajing Securities, said that Chinese biotechs can be vulnerable to fluctuations in revenue from licensing agreements, which are tied to achieving specific milestones in clinical trials while also funding much of their growing research and development spending.
Hengrui’s international deals have highlighted a concern for Chinese pharma companies seeking to get international approval for new drugs. The US Food and Drug Administration has repeatedly rejected one of Hengrui’s cancer drugs, citing questions about quality control at its manufacturing sites.
Akeso
Guangdong-based Akeso has captured investor attention for its advances in oncology. Ren described the company as “very good at drug discovery and clinical execution”.
Akeso’s lead therapy, ivonescimab, has been greeted with a mix of enthusiasm and scepticism. The bispecific antibody, which targets two cancer pathways, was approved in China last year — the first therapy of its kind to reach the market — sending its shares up more than 90 per cent in 2025. It was initially approved for use in patients with lung cancer who had already undergone treatment, but its application has since been expanded.
In Chinese trials, ivonescimab outperformed Merck’s Keytruda, the world’s best-selling cancer drug, in preventing tumour progression. “There’s a bull-bear debate between people who believe the Chinese data,” said Loncar. “The potential value of this drug is exponential if it is better than the greatest cancer drug of all time.”
However, in a recent global Phase 3 trial, the therapy failed to replicate those results among North American and European patients, with their cancers returning and advancing faster than in Chinese studies. Analysts say it is too early to judge the drug’s global prospects, as Akeso expands testing for other cancers, including colorectal.
While ivonescimab remains the company’s key drug, Akeso also has a broad pipeline across oncology, autoimmune and metabolic diseases. Analysts highlight its growing focus on antibody-drug conjugates (ADCs) — these deliver targeted chemotherapy directly to cancer cells, minimising damage to healthy tissue.
ADCs have become a focus for China’s biotech sector, playing to its strengths in large-scale lab experimentation. “Developing ADCs involves huge amounts of trial and error,” said Zhao. “It’s an engineering problem — and China has a surplus of engineers.”
Legend Biotech
Legend Biotech is based in New Jersey, but has its roots in China and its biggest shareholder remains China-based GenScript. Legend is among China’s leading gene therapy developers. It rose to prominence after a 2017 partnership with Johnson & Johnson to develop a CAR-T therapy for multiple myeloma, a type of blood cancer. The landmark deal came in for some criticism at the time as a risky bet on China’s then nascent biotech industry, but it helped to spark the current wave of licensing deals.
While US groups have poured resources into gene therapy, many have been held back by soaring costs and regulatory hurdles in their home market. In China, by contrast, regulators have supported the field by allowing earlier human trials and more flexibility in how they are designed.
Legend’s Carvykti therapy with J&J — which aims to re-engineer a patient’s immune cells to target and destroy cancer — was approved last year in China for patients with myeloma that had recurred after previous treatments.
But while Chinese regulators have nurtured innovation, low domestic drug prices have forced companies to look abroad to recoup investment. “One huge disappointment has been that the commercial sales for China’s drug sector have never bloomed,” said Loncar.
“It has some of the lowest prices for drugs in the world. Nearly all of the success of these companies — and their booming stock prices — hinges on their ability to succeed in the US.”
US government debt burden on track to overtake Italy’s, IMF figures show
Washington to become more indebted than European nations known for fragile public finances for first time this century
The US government’s debt burden is on track to exceed levels in both Italy and Greece for the first time this century, according to IMF forecasts, underscoring the parlous state of America’s public finances.
General government gross debt in the US will rise by more than 20 percentage points from now to reach 143.4 per cent of the country’s GDP by the final year of the decade, IMF forecasts show, exceeding previous records set after the pandemic.
That comes as the IMF estimates that the US budget deficit will hover above 7 per cent of GDP every year until 2030 — the highest of any rich nation tracked by the fund for this year and the rest of the decade.
Italy and Greece have long been highlighted by economists for their fragile public finances. Both countries were at the heart of the euro area sovereign debt crisis of 2010-2012, with Greece requiring a bailout and restructuring overseen by the IMF and EU.
But government debt burdens in both European countries are projected to be on a downward trajectory at the end of the decade as they keep a tight grip on their budget deficits.
By contrast, the US debt-to-GDP ratio will still be rising in 2030, according to the IMF data released this month, with the Congressional Budget Office expecting it to increase for decades thereafter.
“It is a symbolic moment, and according to the CBO the projections are for US debt to carry on rising — that is the impact of running perpetual deficits,” said Mahmood Pradhan, head of global macro at the Amundi Investment Institute.
“But Italy has a weaker growth outlook than the US, so this should not be read as meaning Italy is out of the woods.”
As it boasts the global reserve currency, the US has much more borrowing capacity than European nations.
However, said James Knightley, US economist at ING, “many US politicians and investors look down somewhat on Europe and its slow growth and struggling economies, but when you have metrics like this, the conversation changes”.
The US federal deficit expanded rapidly under the Biden administration, despite unemployment hovering around record lows. The IMF projections show officials believe the Trump administration is doing little to address the problem.
Joe Lavorgna, economic counsellor to US Treasury secretary Scott Bessent, said this month that the Trump administration had made progress in cutting spending and raising revenues through tariffs on US imports.
“What people are missing is the fact that much of the improvement in this year’s fiscal deficit has happened from April onwards,” Lavorgna told the Financial Times.
US gross general government debt has remained below the levels of both Italy and Greece since at least the beginning of the millennium, according to the IMF data. The gauge is a broad measure of indebtedness that includes both central and local government.
An alternative measure — net government debt, which offsets financial assets — shows the US still around 10 percentage points below Italian levels of indebtedness at the end of the decade.
Joe Gagnon, of the Peterson Institute think-tank, said the net measure was a better read on the US’s debt burden, as it closely reflects the portion of debt that investors need to hold. “But this net measure is rising too,” he said.
By contrast, the IMF expects Italy’s net debt burden to be falling from 2028 onwards. It did not give net debt projections for Greece.
Italy has long struggled to curb its debt load because of feeble GDP growth rates, with the IMF forecasting growth of just 0.5 per cent this year and 0.8 per cent in 2026.
Still, Italian Prime Minister Giorgia Meloni’s government has won plaudits from foreign investors for its efforts to pare back Rome’s budget deficit. This year, Italy is forecast to end the year with a primary surplus of 0.9 per cent of GDP, higher than its initially forecast 0.5 per cent.
Rome expects a fiscal deficit — which was 8.1 per cent of GDP in 2022, the year Meloni took office — of 3 per cent of GDP this year, which would allow Italy to exit the EU’s excess deficit proceedings a year earlier than planned.
“There is a continuing, cautious approach to fiscal policy,” said Filippo Taddei, senior European economist at Goldman Sachs.
DBRS Morningstar upgraded Italy’s sovereign rating from “A low” to “BBB high” this month. Analysts say Italy’s quest to strengthen its public finances has been buoyed by its access to more than €200bn in funds from the EU’s pandemic recovery programme.
Carlo Capuano, deputy head of the sovereign rating team at Scope Ratings, said Italy also benefited from a pick-up in the labour market and higher tax collection, buoyed in part by growing use of digital payments.
By contrast, Gagnon said the US political situation made it difficult to see how the country’s yawning deficits could be narrowed, no matter who was in power.
“Democrats don’t want to cut spending and Republicans don’t want to raise taxes,” said Gagnon. “They both want to cling on to that. I don’t know when that dynamic will change.”
Maury Obstfeld, a former IMF chief economist who is now a professor at Berkeley, said any forecast that the US’s fiscal position was sustainable “has to be based on wishful thinking about future US productivity growth, tariff revenue, demographics or interest rates, or possibly all of the above”.
Netflix’s revenue growth rate significantly outpaces major streaming rivals.
Netflix generates billions in cash annually, valuing the company at $470 billion.
Netflix’s ad business is on track to more than double revenue this year.
Netflix’s quarterly earnings reports have lately lacked the drama that once made the updates a must-watch for investors and reporters. For the past eight quarters, the video streamer has routinely posted quarterly revenue growth averaging around 15%—hardly hypergrowth but a big improvement on the 2% to 4% growth rate it recorded between late 2022 and late 2023. Nowadays the results are so predictable that it would be easy to tune them out.
But Netflix is achieving something rarely seen in business. Since last fall, it has been growing at roughly double—or more—the rate of two of its biggest rivals in streaming, Walt Disney Co. and Warner Bros. Discovery, even though those companies have fewer subscribers and have been in the market for less time. As the biggest streaming firm, Netflix would be expected to grow more slowly than smaller rivals. (Other major rivals, such as Amazon Prime Video and YouTube, don’t disclose total revenue, making comparison impossible.)
That reality explains a lot of what is going on in streaming nowadays, including why WBD is now on the sales block—if the company had grown faster, its stock price would be higher and it wouldn’t be so vulnerable to a hostile takeover. It also explains Netflix’s premium stock price, which values the company at close to $470 billion, more than twice Disney’s valuation and several times WBD’s.
It also puts pressure on whoever will succeed Disney CEO Bob Iger, who is due to step down at the end of 2026. While both WBD and Disney have managed to turn their streaming operations into profitable businesses, that’s been at the cost of growth. Netflix, in turn, has sharply accelerated its growth even while continuing to churn out billions in cash every year. Last week, for instance, Netflix reported 17% higher revenue for the third quarter, while free cash flow rose 21% from a year-earlier to $2.7 billion.
In raw numbers, WBD’s streaming operations are generating around $2.8 billion a quarter in revenue, while Netflix’s quarterly revenue is running above $11 billion. Disney is in between (see accompanying chart). At its recent growth rate, Netflix is adding around $1.5 billion every quarter compared to the year earlier, while WBD is adding a couple of hundred million dollars. (Disney and WBD haven’t yet reported for the September quarter, although a report last month by Wall Street research firm MoffettNathanson estimated neither would show much improvement in their growth rate for the second half of 2025.)
At the current growth rates, Netflix is only going to get further ahead in streaming. The more cash it generates, the more money it has to spend on programming.
What’s driving the diverging growth rates of Netflix compared with Disney and WBD? Over the past 18 months, Netflix’s quarterly disclosures have attributed its revenue increases to adding subscribers and raising prices. In the past two quarters, it has also mentioned growth in advertising. That’s a sign that the ad business, which Netflix only entered in late 2022, is now becoming significant. But the company doesn’t disclose enough detail to show how much of the growth comes from either subscriber increases or advertising.
Adding subscribers should be harder for Netflix, which finished last year with 301 million global subscribers and therefore should be closer to saturating the market. Disney and WBD are much smaller—156 million for Disney and 125.7 million for WBD as of the end of June. (Netflix stopped disclosing subscriber numbers at the end of 2024 and Disney has said it will stop disclosing them in the fiscal year which just began).
Disney’s revenue growth slowed sharply in the past 12 months, along with its subscriber growth. Some of that is an accounting adjustment due to its transfer of its Indian streaming service into a joint venture in late 2024, to be sure. But its subscriber growth rate elsewhere in the world has slowed at the same time. That is likely a result of the price increases that have driven what revenue growth it has reported.
Meanwhile, WBD has done better at adding subscribers as it has expanded into new global markets. But the average price it can charge in most of these markets is a third of what it charges in the U.S., which means its revenue growth lags its subscriber growth.
The average revenue it gets from subscribers in the U.S., meanwhile, has been dampened by bundling deals with other companies where WBD gets less per subscriber than it would on its own.
Like Disney and Netflix, Warner has relied on regular price increases to juice revenue—and WBD CEO David Zaslav has indicated that’s likely to continue. He told a recent investor conference, sponsored by Goldman Sachs, that consumers nowadays are paying “dramatically less” for content than they were a decade ago. As a result, he argued that “over time, there’s a real opportunity…to raise price.”
Netflix may be doing a better job of balancing price increases with growth. Its lowest tier of service—the one that carries ads—is cheaper than rivals’ comparable services, according to MoffettNathanson’s recent report. That tier is only $7.99 a month, $2 or $3 a month less than most other services. That enables it to more easily draw in price-sensitive consumers.
Its top tier, on the other hand, is pricier than its rivals’. As MoffettNathanson analyst Robert Fishman said in last month’s report, “This pricing structure allows Netflix to capture the highest-value, least price-sensitive customers.”
Certainly the gap is closing at the top end. WBD announced this week that HBO Max’s top price would rise to $23 a month, $2 below Netflix’s top tier.
Both Disney and WBD might do better in streaming growth next year. In 2026 WBD is expanding HBO Max into the U.K., Germany and Italy, three markets where its pricing should be closer to the U.S. market. Disney, meanwhile, should be able to jump-start its streaming growth in the next year thanks to the recent launch of its ESPN streaming service.
Netflix, though, has its own growth driver, one that might accelerate—its ad business. The company hasn’t yet revealed how much money it is making from selling ad space, although co-CEO Greg Peters said on the third-quarter earnings call on Tuesday that the company was “on track to more than double ad revenue this year.” He noted, though, that the increase was “off a small base relative to the size of our subscription revenue.”
One issue that Netflix has had is that the proportion of its audience seeing ads may not be as high as it is for Amazon Prime Video, which introduced ads by putting them into every subscriber’s service unless they opted out. Over time, however, the fact that Netflix’s ad tier is cheaper than its rivals’ means it is likely to attract more subscribers. And because of Netflix’s popularity, advertisers will want to buy spots on the service.
Still, advertisers have some complaints about Netflix’s ad offering. One ad agency executive, David Nyurenberg, a senior vice president of digital at ad agency InterMedia Advertising, said Netflix differs from other streaming services in its unwillingness to give advertisers the data they need to assess whether its ads drive sales. Other streaming services give advertisers data allowing them to determine whether people who are shown an ad later make a purchase.
“I would love to be on Netflix, at least test it out, but the performance measurement makes it a nonstarter,” Nyurenberg said, adding that advertising on Netflix is also relatively expensive in terms of cost per thousand impressions, which also made the streaming service a tough sell for his clients.
Another ad buyer also said they invested first in Amazon Prime Video before spending money on Netflix, because Amazon had more experience in advertising.
Vinod Khosla Talks AI, Power and Why Businesses Struggle With New AI Services
The Takeaway
Enterprise AI adoption struggles due to unqualified internal staff.
AI product margins are expected to increase healthily into the 2030s.
AI’s circular financing deals pose no systemic risk, Khosla says.
Vinod Khosla thinks companies which say they’re struggling to get a return on their spending on artificial-intelligence services may be doing it wrong. In an interview with The Information’s TITV this past week, he said “most enterprises who are executing AI are doing it with their people who are not qualified to execute.” In contrast, when businesses hire an AI specialist firm to help them, “it goes swimmingly well.”
Khosla is one of the most experienced venture capitalists in Silicon Valley, whose firm was a very early investor in OpenAI, among many others. That makes his take on the state of play in AI worth listening to. In the TITV interview, he talked about current developments in AI, the risks inherent in the AI ecosystem right now, power generation and other topics.
Below is a lightly edited transcript of the interview. The full video is here.
Akash Pasricha: Our subscribers at The Information know your career very well. They will know that you are a steadfast optimist, and one of the guiding mantras in your career has been that skeptics never did the impossible. It feels like right now, even for people who are very bullish on AI, the technology, there is enough reason to be at least somewhat skeptical on AI, the business. Is there anything that you are skeptical about right now?
Vinod Khosla: Well, to start with, I think if you’re going to do something significant, you’re going to have to be an optimist. But not only just an optimist, a knowledgeable optimist. There are too many skeptics who live in the past and will extrapolate the past to invent the future. That’s a bad thing to do. Inventing the future you want is the right way, and optimism is key to that.
Having said that, I think the one thing I don’t know and there’s things I can reasonably project and things I can be uncertain about. The thing I’m uncertain about is timing of various breakthroughs that we expect to see in the next two years, five years, ten years. Almost certainly it’s easier to predict AI in 2035 and probably 2030, than it is to predict it today and what will happen next year. So I’m not skeptical, but I’m less certain about the pace of development of certain capabilities that are essential to expanding use of AI.
Right. Well, let’s talk about the current state of play. One of the topics that has been talked about a lot in our coverage are the margins that a lot of these companies have on their AI products, whether it’s cloud services or even AI application companies. And one of the challenges that we’ve highlighted is that margins are slim right now. How do you think this goes in the long run?
Vinod Khosla: I think the way to look at margins is when AI is a substantial business. Yes, it’s [a] large business, tens of billions of dollars, but it’s not hundreds of billions of dollars. And so the question to ask is where will margins stabilize? I think some of it depends on R&D breakthroughs by the various companies and knowing who gets what breakthrough, which is, as I said, uncertain with respect to who and timing. I’m pretty optimistic about what OpenAI is doing. I do think if you provide value and you have differentiated models, you will have good margins. Look, at the pace at which pricing is declining, both in supplying inferencing to the AI companies and the price at which things are sold, it’s hard to predict the dynamics. But it will settle down [with] good returns on capital invested. So I’d be surprised if we don’t have healthy increasing margins well into the early 2030s.
And which of the two levers do you think is more likely to get pulled on most? The idea that costs will go down for these chips that are very expensive right now. Or on the flip side, we had one of your founders on the show, Amjad Masad from Replit. He talked about this idea that the models might not get cheaper, but rather he’s looking to actually increase his prices to widen his margin. So which one of those two levers do you think is most likely to widen margins in the next couple of years?
Vinod Khosla: So there’s a number of ways in which cost of supplying inference will go down. One, chips will become more economical in per inference. I do think the algorithms will get much better, which means the software will get better in the amount of compute needed per inference. So there’s two vectors. You can make the algorithms 10X more efficient over time. I think that will happen. The cost of chips will go down for a given number of inferences. So both those are vectors for cost reduction.
The question of pricing on the input side, what customers will pay, will be a function of value. If Amjad—and I love that company and what Amjad is doing—if he keeps adding more and more value, he will be able to charge more. So that’s a question of value addition. And that’s where there’s lots of headroom. You know when you’re kicking something that costs, say, a professional, an accounting professional, for example, or a design product designer, and you’re paying them $100 to $300 an hour and your cost is $1 to $3 an hour, you have lots of pricing room if you can provide more complete solutions. So I do think, price per hour of worker-time equivalent will increase pretty dramatically. Also as well, as well the decline in costs per inference. This business will be healthy starting, let’s say, 2030 and beyond, when scale data centers will be in operation.
How do you square that with this matter of buyers of enterprise software—businesses in general—they are struggling to find the value right now, at least on an ROI basis, for much of this AI software. The reason I’m asking the question is because we’re talking about raising prices. They’re not seeing the value right now. They’re very slow to adopt this AI software. So how do you square those two realities right now?
Vinod Khosla: So you have to look at a couple of factors. One, where is the value being provided great. In software development it’s absolutely great value. So companies like Replit and Cognition, which we are both investors in, Cursor included, all growing very, very rapidly, because they’re adding real value. So there are functions in which the product isn’t complete or mature. And so you need almost so much hand-holding, the economics break down. I would say another factor that most people haven’t considered, most enterprises who are executing AI are doing it with their people who are not qualified to execute. It’s like saying, hey, we have a race car and Joe Blow can go drive it, and he’s not going to get most of that race car.
So they need to hire different people?
I think they need to take a very different approach. They take an online IT software person, say build an agent for me and hope it works great. It’s not the way it’s working. So generally if you take a company like Distyl, which is in our portfolio, if they execute a project for a large Fortune 500 company, it goes swimmingly well. If in-house people in the same company executed, it goes very poorly. So, and each of these, even the in-house people will get better over time, so the third or fourth generation of execution will do better because their people will get trained, but they’re really not qualified to operate in this area. While the AI native companies are able to get real value, and they’re even able to pick the projects that will be valuable and the projects that are more experimental.
I want to pivot to talking a little bit about some of the circular deals that we’ve seen happening in the AI sector. I mean, Nvidia, for instance, has been at the center of some of these circular financing arrangements. Do these concern you at all?
Vinod Khosla: Well, they don’t, they do and they don’t. It’s hard to tell what the details are behind each of these deals. If Nvidia is financing customers to buy their chips, that could be perfectly reasonable. Look, General Motors finances its cars too, when a consumer buys it, it’s just a regularized business. The question is hidden in the contracts that are mostly not publicly available: Who’s taking on what part of the risk? Is it an enterprise? Are you saying a customer you’re financing is viable or not? Are you saying the risk is the customers’ or their customers if they’ve done a contract with somebody else to buy X million dollars of inferencing? So where’s the risk hidden? Well, in fact is the key question before one can opine. And frankly, most of those risks and who takes what risk is in the fine print and not visible to almost anybody outside?
I take your point on the fine print. I guess what I’m sort of trying to assess here is the systemic risk in the circular financing at large. And of course, I take your point about, for example, General Motors financing the purchase of a car. But, you know, in a world where you have a chip company investing in OpenAI, OpenAI buying cloud compute from Oracle, Oracle buying chips from Nvidia. That is a sort of circular loop that I think a lot of people have sort of raised flags about. Does that whole cycle not raise any flags for you or cause any concern?
Vinod Khosla: Well, I would say I don’t care. I don’t care for the following reason. If Nvidia is taking a bad credit risk, that’s their problem, right? But if Nvidia loses $50 or $100 billion, does it kill the company? Probably not. And I suspect [Nvidia CEO] Jensen [Huang] is pretty smart about what credit risks he’s taking, with which customer. Is Oracle taking a larger risk? Possibly. It depends on the details of the contract between Nvidia and Oracle and Oracle and OpenAI or other people buying their cloud service.
So you have to think of it as traditional business and say where does the risk lie? If Oracle goes under, for example, because they took the risk of $100 billion spend and then get that back, then it’s their problem if Oracle disappears from the scene. Do I care? No. I hope they don’t. I think the ecosystem will be healthier, but people are selectively taking risk. You know, Coreweave’s taking a lot of risks with the money that belongs to certain lenders, to Coreweave. Do the lenders have risk? I don’t know. Does Coreweave have risk? And Oracle’s doing the same thing, it’s taking risk. But I don’t know the details of these contracts.
But broadly speaking the level of risk that this has has become sort of ubiquitous throughout this AI ecosystem that doesn’t concern you at all?
Vinod Khosla: I would say the fundamental notion of will there be more demand for API inference calls doesn’t concern me at all. The fundamental is how many inferencing calls we’ll see in 2030 and 2035. That generally doesn’t concern me because I believe AI will add a lot of value. Look, US economy’s $15 trillion of labor alone—just labor costs in the US economy $15 trillion. If you could replace 5 trillion of that, there’s plenty of room for inferencing to be paid for, if you can do that. So again I say the fundamental is, is there a demand for AI inferencing the next five years and next ten years, I’m not worried about that. has learned how clever a contract and who takes on risk if demand is slower or faster to emerge? … I’m not responsible for a lender financing a data center. If they fail, their problem, not mine. I care about the innovation ecosystem that drives more API calls in AI.
Who do you think has the best chance of challenging Nvidia?
Vinod Khosla: Well, obviously AMD is trying things, Arm is trying things, Broadcom is trying things. Nvidia is in an unenviable position because they have so many different things they can do in parallel because the cash flow they have right now. Do I know all of Jensen’s plans inside and how many different things he’s trying? No. In fact, nobody outside really knows when he’s going to announce what my bet is he has a pretty precise roadmap to 2030 and beyond. So hard to say.
So there’s no one company that you are really sort of putting your eggs in here in terms of… I think this company is closest on Nvidia’s tail right now.
Vinod Khosla: Well, AMD is doing pretty well by signing, looking at their deals. Broadcom is doing pretty well. But are they going to grab majority share and be larger than Nvidia. I wouldn’t expect that today. Can they take reasonable shares, especially at slightly lower margins? Yes.
And what about all these chip startups that are popping up.
Vinod Khosla: Well I’ve seen a lot of specialized chip startups that do one thing. You can run your whole model locally in your shop. Well that’s that’s a market. It’s not as large as the data center inferencing market. So there’s many submarkets that the chip startups can do ok in.But I haven’t seen the chip startup that could completely blow everybody away. Now, if you have a sudden breakthrough in photonic chips that can do multiply, accumulate inferencing functions, and cut the power consumption by 70% for inferencing, that’s entirely possible, even likely sometime in the next five years.
And I hope some of those show up because it changed the power equation, how much power we need for AI. Frankly, most of the data center investment can be repurposed with a new kind of chip that gets slotted in. Photonics is pretty promising. I suspect digital semiconductor chips are going to be hard to beat Nvidia at, in a massive way. In specialized segments of the market, you can beat them but generally you’d have to have a radical breakthrough in technology. Photonics is one of them. There’s a few others, but I see the most promising candidate for an alternative to Nvidia becoming from photonics if one can scale it, and photonics typically is hard to scale.
I want to close by talking about the energy side of of the AI equation. You, of course, have been an energy for a long time. You’ve made big bets on fusion among many other technologies. And the question that I want to help get your perspective on is this idea that a lot of these energy bets are going to still take years to scale? Meanwhile, the energy demands that AI will need to satisfy, those demands are right now. And so help us reconcile this idea that we need power now, and it’s going to still take several years for many of these new energy technologies to scale.
Vinod Khosla: So the simplest way, very, very short term, to address electricity demand in the country is pricing. This is why the economics of marketplaces work. Prices will go up some, as we consume more pricing for data centers. Data centers themselves can be aggressive sources of power management. You can consume at certain times of the day for training runs and other times of the day for inferencing. You can dial up the performance or dial down the performance. So data center input of electricity itself is a variable that’ll probably be part of electricity trading.
There’s some good startups in that area.
I think there’s a short term solution which I think is super hot geothermal. I think we can get to gigawatt scale. If you imagine a couple of extra gigawatts of demand emerging every year. Some of it will be met through pricing energy appropriately. Some of it will be met through shorter term projects. Geothermal is one that’s much shorter term than, say, fusion, fusion’s a possibility but fusion to me will take the longest…but we’ll have an array of factors. We’ll have more natural gas turbines coming on. We do have companies where you can start with natural gas and switch to hydrogen whenever the economics warrant it. So there’s a number of ways to adjust, but it is a non-trivial problem.
And so all these data centers that companies like OpenAI are springing up very quickly. I mean, the simple question I wanted to get your take on is, do we have enough power for these facilities?
Vinod Khosla: Well, first thing to keep in mind, it takes a couple of years to build a data center. There’s a few hacks, but fundamentally, if you’re trying to add a gigawatt of data center, which is about $30 billion of spend, it’ll take a couple of years. I don’t think that’s a six month or 12 month or even an 18 month project.
Then there’s demand-based electricity consumption as a tool. And then I think things like geothermal and other technologies will come along. Some of them will be natural gas fired. I hope there’s no more coal facilities. Mainspring installs capacity for data centers that can switch seamlessly from natural gas to hydrogen when you want to go clean, so you can decide how much you want to pay for power and what carbon reduction you want, and increase the carbon reduction over time. So that’s one solution. All I’m saying is there’s an array of tools …not non-trivial. This will be a serious issue. And policy is trying to address it. But I do think there’s many solutions.
Last question for you. You know, you wrote this op ed for us a couple months ago in The Information, about the bonkers valuations in AI right now. And one of the things that you mentioned in the piece was that you think that venture capital as an asset class is likely to shrink over the coming ten years. And I wanted to ask you what the repercussions of that are in your mind. The obvious one that I thought was perhaps startups might actually get better because there’s less capital to go around and the better ones would get picked. Is that the main repercussion of this, or what are the other repercussions that you see happening?
Vinod Khosla : I think that op ed was misinterpreted a little bit. What I said was AI valuations in general are bonkers. For the best companies they’re not bonkers and if as a venture capitalist, you have access to those two, three, 4% of the startups, there will be huge wins. You will do well. You will have great returns. The people who are plowing money in without having special access to these opportunities for whatever reason, will suffer in venture capital as a class, I think broadly, for funds raised in ‘24 and ‘25 will have decreasing returns, returns lower because they’re not getting access to good deals early. They’re paying higher prices much later. Some of the robotics valuations are getting bonkers. I would venture to guess 95% of those startups will lose money.
So the take home here is we need special access to deals essentially to be successful?
Vinod Khosla: It’s more than that. I think I like to say most AI startups will lose money, but more money will be made than lost. That means it’ll be highly asymmetric: 2 or 3% of the startups will account for 85-90% of the valuation by 2035 of market cap of companies. So that asymmetry, which has generally been true in venture capital, but will be significantly more asymmetric in AI because of this valuation wave, I think is the reason we will see average returns decline and the best returns for the top firms will do okay, will do well, in fact, because AI is such a large opportunity.
It’s Not Just Rich Countries. Tech’s Trillion-Dollar Bet on AI Is Everywhere.
As part of ‘AI decolonization,’ developing nations push Silicon Valley to build locally
Indonesia is emerging as an AI hub, with its market projected to grow 30% annually to $2.4 billion over the next five years.
The concept of AI decolonization drives developing nations to require local data processing, fostering domestic data centers.
Telecom provider Indosat plans to increase its AI-enabled server power to 500 megawatts by 2028, a significant investment in Indonesia’s AI infrastructure.
JAKARTA, Indonesia—Tap water isn’t drinkable. Power outages are common. The national average annual wage is $2,200.
Yet rising on Jakarta’s outskirts are giant, windowless buildings packed inside with Nvidia’s latest artificial-intelligence chips. They mark Indonesia’s surprising rise as an AI hot spot, a market estimated to grow 30% annually over the next five years to $2.4 billion.
The multitrillion-dollar spending spree on AI has spread to the developing world. It is driven in part by a philosophy known in some academic circles as AI decolonization.
The idea is simple. Foreign powers once extracted resources such as oil from colonies, offering minimal benefits to the locals. Today, developing nations aim to ensure that the AI boom enriches more than just Silicon Valley.
Regulations effectively require tech companies such as Google and Meta to process local data domestically. That pushes companies to build or rent data facilities onshore instead of relying on global infrastructure. These investments add up to billions of dollars and create jobs that foster national talent, or so developing nations hope.
“The model is shaping up where it has to be more fair to the country,” said Vikram Sinha, chief executive of Indosat Ooredoo Hutchison, a partially government-owned Indonesian telecom provider building data-center projects across the country.
AI decolonization is a twist on data sovereignty, a concept that gained traction after Edward Snowden revealed that American tech companies cooperated with U.S. government surveillance of foreign leaders. The European Union in 2018 pioneered data-protection laws that other nations have since mimicked.
Regulations vary by country and industry, but the principle is this: If a developing-nation bank wants an American tech giant to store customer data and analyze it with AI, the bank must hire a company with domestically located servers.
Combined with genuine business demand for AI computing, such regulations have made some less-wealthy nations beneficiaries of the AI boom, including in South and Southeast Asia.
Google recently committed $15 billion over five years to build an AI center in India. Amazon Web Services is investing $5 billion to increase infrastructure in Thailand.
Data centers in Indonesia earned $374 million in revenue last year, said research firm Mordor Intelligence, a figure estimated to sextuple by 2030.
Betting big on this new sector is Indosat, Indonesia’s second-largest telecom provider. Its CEO, Sinha, was emboldened by Nvidia Chief Executive Jensen Huang, who championed “sovereign AI” during a visit to Jakarta last year.
“No country can afford to have its natural resource—the data of its people—be extracted, transformed into intelligence and then imported back into the country,” Huang said at an Indosat event.
Sinha had a chance to ask Huang for advice during a dinner. He recalled the answer: “If you believe in something, go all in.” To Sinha, that belief is AI.
Indosat plans to increase its AI-enabled server power to 500 megawatts by 2028 from 20 megawatts currently, in data centers with backup electricity. By comparison, said research firm DC Byte, the American hub of Virginia has about 38,000 megawatts of capacity live or in planning.
The industry rule of thumb is $10 million a megawatt—so a $5 billion investment for Sinha. He thinks Indosat’s Indonesian economics can reduce the cost to 70% of that.
Potential clients include Indonesian customers and American tech giants, who outside the U.S. often use local partners instead of building their own facilities.
Sinha said Indonesia’s data-center boom would supercharge the economy and boost the tech talent pool. Other experts are less certain.
After construction, each megawatt of a data center supports about only one permanent job, said James Murphy, Asia-Pacific managing director of DC Byte. And even if built in developing nations, many data centers remain foreign-owned.
“The ultimate beneficiaries are the large tech companies and their investors,” said Govand Azeez, a political economist at Sydney’s Macquarie University who researches tech sovereignty in Southeast Asia.
