Latest / DX Today | No-Hype Podcast & News About AI & DX / AI News Today: OpenAI Revenue Comes In $20B Short

Transcript

AI News Today: OpenAI Revenue Comes In $20B Short · DX Today | No-Hype Podcast & News About AI & DX

• 0:00
  SPEAKER_05: It's Friday, October 9, 2026.
• 0:03
  SPEAKER_05: You're listening to the DX Today AI Daily Brief.
• 0:07
  SPEAKER_05: Today, OpenAI's revenue comes in roughly$20 billion below what investors were led to expect.
• 0:15
  SPEAKER_05: Google Cloud unveils a single AI agent for work, and the White House lines up$2.4 billion in computing pledges for science.
• 0:25
  SPEAKER_05: Let's get into it.
• 0:26
  SPEAKER_00: OpenAI has told investors its annualized revenue at the end of September was close to$50 billion.
• 0:34
  SPEAKER_00: That is about$20 billion less than the figure investors had been led to expect, according to the Financial Times, with Reuters confirming the number through a person familiar with the matter.
• 0:44
  SPEAKER_00: The gap reportedly came from an earlier attempt to compare OpenAI's numbers directly with rival Anthropic, which counts some sales through cloud partners differently.
• 0:53
  SPEAKER_00: The report rattled markets on Thursday, shares of NVIDIA, Oracle, Core Weave, and MemoryMaker Micron fell, and tech stocks in Asia struggled overnight.
• 1:04
  SPEAKER_00: Even at$50 billion, OpenAI's run rate has still grown by more than 70%.
• 1:09
  SPEAKER_00: But the episode raises fresh questions about how AI companies present their revenue to the investors funding them.
• 1:15
  SPEAKER_02: That news hit one backer hard.
• 1:17
  SPEAKER_02: Softbank, among OpenAI's largest investors, saw its shares slide on the revenue report.
• 1:25
  SPEAKER_02: At the same time, the Financial Times reports that founder Masayoshi-sun is seeking up to$100 billion from Gulf investors, including contacts in the United Arab Emirates.
• 1:38
  SPEAKER_02: The money would go into a fund that buys companies and improves their operations using artificial intelligence.
• 1:46
  SPEAKER_02: The talks are not a done deal, and the paper said success is not guaranteed.
• 1:52
  SPEAKER_02: The pitch comes right after Softbank completed its$30 billion investment in Open AI and sold$11.1 billion of bonds last month, one of the largest high-yield bond sales on record.
• 2:08
  SPEAKER_02: The move shows how far Sun is willing to stretch SoftBank's balance sheet to stay at the center of the AI boom.
• 2:16
  SPEAKER_04: Now a new coworker from Google.
• 2:18
  SPEAKER_04: Google Cloud has introduced the Gemini agent, which it calls a single universal agent for work.
• 2:24
  SPEAKER_04: Chief Executive Thomas Curion unveiled it at the company's Gemini at work event on Thursday.
• 2:30
  SPEAKER_04: Instead of answering one prompt at a time, the agent takes a goal, plans the steps, and returns finished work.
• 2:37
  SPEAKER_04: It can research questions, draft documents, create images and media, and write and run code, all grounded in a company's own data.
• 2:46
  SPEAKER_04: It works inside tools like Gmail, Docs, Slack, and Microsoft 365.
• 2:52
  SPEAKER_04: The product is still in early access.
• 2:54
  SPEAKER_04: The launch puts Google in direct competition with OpenAI, Anthropic, and Microsoft for the business of handing everyday office work to AI agents, the market every major lab now sees as its next big revenue stream.
• 3:07
  SPEAKER_04: Turning to China's agent startups.
• 3:10
  SPEAKER_01: Manus, the Chinese AI agent startup, has raised more than$500 million in its first funding round as an independent company again.
• 3:19
  SPEAKER_01: Its parent, Butterfly Effect, announced the round on Thursday.
• 3:23
  SPEAKER_01: It was jointly led by Boyou Capital and IDG Capital, with existing backers including Tencent, taking part.
• 3:32
  SPEAKER_01: The raise follows the collapse of Meta's planned acquisition of MANIS, a deal valued at more than$2 billion, which was unwound after the Chinese government blocked it.
• 3:42
  SPEAKER_01: MANIS builds general purpose agents that carry out multi-step tasks online on a user's behalf.
• 3:49
  SPEAKER_01: The new money gives it room to compete on its own in a crowded field.
• 3:53
  SPEAKER_01: And it signals that investors still see value in Chinese agent technology despite growing friction between Washington and Beijing.
• 4:03
  SPEAKER_03: The Trump administration says 11 technology companies have pledged$2.4 billion in computing power and AI tools to the Genesis mission.
• 4:11
  SPEAKER_03: That is the federal effort to use advanced AI to speed up scientific discovery at Department of Energy Labs and other agencies.
• 4:19
  SPEAKER_03: Partners named in Thursday's White House fact sheet include Nvidia, OpenAI, Anthropic, Amazon, and Google.
• 4:28
  SPEAKER_03: The commitments are mainly compute credits and technology support rather than cash, and they will serve more than 15 federal agencies.
• 4:36
  SPEAKER_03: The announcement was part of a broader science package that also includes new quantum computing funding.
• 4:42
  SPEAKER_03: Critics noted the pledges arrive while the administration has proposed cuts to traditional research budgets, so how much new science the credits actually buy will be worth watching.
• 4:52
  SPEAKER_05: Meanwhile, the public is uneasy.
• 4:54
  SPEAKER_00: Most Americans think artificial intelligence is developing too fast.
• 4:58
  SPEAKER_00: That is the finding of a new poll from the Associated Press and the Norc Center for Public Affairs Research.
• 5:04
  SPEAKER_00: About 64% of adults say AI is moving too quickly.
• 5:08
  SPEAKER_00: Roughly eight in ten say keeping AI under human control and protecting American workers should be important priorities for the federal government.
• 5:16
  SPEAKER_00: Those worries cross party lines.
• 5:18
  SPEAKER_00: The poll also found that neither party has an edge on the issue.
• 5:21
  SPEAKER_00: About four in ten Americans say they trust neither Democrats nor Republicans on APAI.
• 5:26
  SPEAKER_00: The results land as the White House promotes its own AI agenda, and they suggest that public trust, not just technology, may decide how quickly these tools spread.
• 5:37
  SPEAKER_02: Next, the chip wars shift.
• 5:39
  SPEAKER_02: Nvidia plans to invest in D-Matrix, a seven-year-old startup that designs AI inference chips meant to compete with Nvidia's own.
• 5:50
  SPEAKER_02: That's according to the information, citing three people familiar with the deal.
• 5:55
  SPEAKER_02: Terms were not disclosed, and the investment has not been confirmed.
• 6:00
  SPEAKER_02: It would be Nvidia's first investment in an inference chip startup since its$20 billion licensing agreement with Grok last December.
• 6:10
  SPEAKER_02: D-Matrix already works with Nvidia, and its next generation chips are designed to plug into Nvidia's NVILLINK Fusion ecosystem.
• 6:20
  SPEAKER_02: The strategy is clear.
• 6:22
  SPEAKER_02: Rather than fight every challenger, Nvidia is pulling rivals into its own hardware platform, keeping itself at the center of AI data centers even as customers seek alternatives.
• 6:37
  SPEAKER_04: And a deal for packaging.
• 6:38
  SPEAKER_04: Global Foundries has reached a$2 billion agreement to make a key AI chip component for TSMC.
• 6:45
  SPEAKER_04: The deal covers silicon interposers, the thin layer that sits beneath processors and memory, and links them together inside TSMC's CoWS Advanced Packaging.
• 6:55
  SPEAKER_04: That packaging step has been one of the tightest bottlenecks in producing top AI chips.
• 7:01
  SPEAKER_04: Global Foundries will add capacity at its Malta, New York plant.
• 7:05
  SPEAKER_04: The agreement has an initial five-year term, with volume production expected to ramp in the first half of 2028.
• 7:13
  SPEAKER_04: For TSMC, it adds supply outside Taiwan.
• 7:16
  SPEAKER_04: For the United States, it brings a critical piece of the AI supply chain onto American soil, something Washington has pushed for since the chip shortages of recent years.
• 7:26
  SPEAKER_05: Now, who grades the models?
• 7:28
  SPEAKER_01: Arena, the company behind the popular LM Arena leaderboard for ranking AI models, has raised$200 million at a valuation of$3.1 billion.
• 7:39
  SPEAKER_01: The Series B was jointly led by Lightspeed Venture Partners and Kozler Ventures, and it nearly doubles the company's value in about 10 months.
• 7:48
  SPEAKER_01: Arena started as a research project at the University of California Berkeley in 2023.
• 7:54
  SPEAKER_01: Millions of users compare model answers side by side and vote for the better one.
• 8:00
  SPEAKER_01: Alongside the funding, Arena launched an alignment index, which measures how well frontier models match human intent and values, including whether they lie.
• 8:09
  SPEAKER_01: As labs compete on benchmarks, independent scorekeepers like Arena are becoming valuable and influential players in their own right.
• 8:17
  SPEAKER_03: Turning to India's services giants, Hexaware Technologies has signed a multi-year partnership with Anthropic, becoming a preferred partner in the Claude Partner Network.
• 8:27
  SPEAKER_03: The Indian IT services firm will build Anthropic's Claude models into the core of its own AI platforms, including Xerovity and Agent Verse, and help enterprise clients deploy them across software development and IT operations.
• 8:43
  SPEAKER_03: Investors cheered.
• 8:44
  SPEAKER_03: Hexaware shares jumped roughly 11 to 12% in Mumbai trading on Friday.
• 8:49
  SPEAKER_03: The deal is the latest sign that large outsourcing firms see AI model makers less as a threat and more as partners.
• 8:57
  SPEAKER_03: By reselling and integrating models like Claude, they hope to protect their client relationships as automation changes the traditional business of writing and maintaining code.
• 9:07
  SPEAKER_05: Results from another services leader.
• 9:10
  SPEAKER_00: Tata Consultancy Services, India's largest IT services company, reported second quarter results on Thursday.
• 9:18
  SPEAKER_00: Net profit rose about 15% from a year earlier, and revenue grew 11.2%, helped by broad-based growth across international markets.
• 9:26
  SPEAKER_00: The headline for many analysts was AI TCS said its annualized AI revenue has now crossed$3 billion.
• 9:34
  SPEAKER_00: It also announced a deal to take over Best Buy's global capability center in India and turn it into an AI capability center.
• 9:41
  SPEAKER_00: Operating margins dip slightly and missed some estimates, and analysts remain cautious on the stock.
• 9:46
  SPEAKER_00: But TCS says margins should recover as its heavy AI investments taper off and more of its staff move on to paying projects.
• 9:54
  SPEAKER_02: Finally, India Eyes New Rules.
• 9:57
  SPEAKER_02: India plans to publish a consultation paper on regulating artificial intelligence within a month.
• 10:14
  SPEAKER_02: At the same time, the electronics and IT Ministry has started a long-term procurement for computing power, covering an estimated 10,000 AI chips over five years, with an initial tender for 5,000.
• 10:30
  SPEAKER_02: India is trying to balance two goals at once, building its own AI capacity under its national mission, while setting guardrails against misuse.
• 10:41
  SPEAKER_02: The country's approach could shape how one of the world's largest online populations experiences AI.
• 10:48
  SPEAKER_05: That's your briefing for Friday, October the 9th, 2026.
• 10:52
  SPEAKER_05: For DX today, stay curious.