Just when you thought the artificial intelligence model race might slow down, it starts up again double-time.

This week alone brought new models and related services from OpenAI (twice), Meta (also twice), SpaceXAI, Anthropic and a raft of Chinese companies. And Microsoft, which released a flurry of seven models last month, intends to ditch OpenAI’s and Anthropic’s models in some cases in favor of its own to save costs in a variety of its platforms.

But this is today’s world: the token economy, as John Furrier puts it. “The bet is no longer on whether the technology works,” he writes. “It is on how fast an economy can be rebuilt around the assumption that intelligence is cheap — and on who is left owning the pipes, the factories and the knowledge when the price of thinking falls to the price of electricity.”

Not surprisingly, the competition is getting even more fierce, including on price, as Meta vows to undercut other models by a lot. And that competition centers squarely on the enterprise business that’s currently the most lucrative market. Palantir CEO Alex Karp, for instance, late the previous week went off on the AI model providers, which he says are trying to control enterprises’ own data — though it’s just as clear that he’s worried about Google, Microsoft, OpenAI and others deploying their engineers inside enterprises in a bid to copy Palantir’s successful model.

“The unresolved question is who captures the resulting value,” Dave Vellante writes in his latest Breaking Analysis on the upshot of Karp’s outburst. “If the system of intelligence remains outside the frontier labs, companies such as Palantir, Databricks, Microsoft, SAP, Salesforce and ServiceNow will have the opportunity to make models interchangeable and capture the enterprise control point. If the frontier labs use their scale to build, buy or partner into the system of intelligence, they could dominate far more of the enterprise stack than today’s skeptics expect.”

All this swirls as investors start to get a little wary: They’ve whacked a trillion dollars off AI uber-leader Nvidia’s valuation, which stands at its lowest point since early 2019 — close to three years before ChatGPT. Even the Treasury Department is warning in an internal report that an AI bubble could pop. At the RAISE Summit in Paris this week, the bubbliciousness was apparent.

AI demand, which so far is real, is also driving unprecedented demand for even more AI chips than Nvidia can supply, such as those from SambaNova, which just raised a billion dollars at an $11 billion valuation, and the memory chips needed to supply all the data enterprises want the models to crunch. Korea’s SK hynix just went public again, this time in the U.S., raising $26.5 billion Thursday evening to make all the high-bandwidth memory chips it can.

Meantime, Samsung’s minting money like never before, forecasting that memory chip demand will multiply profits nearly 20-fold when it reports results later this month. And not to be left behind, Broadcom inked a deal with Apple, which pledged to buy $30 billion of its U.S.-made chips through 2031.

Late-breaking Friday: Apple sued OpenAI for allegedly stealing trade secrets from former Apple employees, and it’s not holding back.

Check out the SiliconANGLE 100, our new index of the top 100 publicly held enterprise and emerging tech companies, powered by QuoteMedia. It’s at the top right of the homepage navigation bar. You can get live stock prices and other data as well as articles from SiliconANGLE and other publications. It’s part of our goal to become the nexus for enterprise news and information.

Here’s all of this week’s enterprise and emerging tech news and analysis from SiliconANGLE and beyond:

AI and data: Model mania

Analysis, opinion and food for thought

The token economy: The state of AI mid-2026

ICYMI: Breaking Analysis: Alex Karp, frontier models and the real fight for Enterprise AI And another view on Karp’s comments: When the sovereign AI diagnosis goes prime time

TheCUBE Research finds Oracle’s AI database move could unlock bigger multicloud returns

Big Tech has suddenly flipped on the AI jobs wipeout scenario (per the Wall Street Journal) “They may have noticed that the labor market is genuinely not changing (i.e., imploding) as rapidly as they expected,” said David Autor, a professor of economics at the Massachusetts Institute of Technology. “They may have realized it was simply bad business to say that your great new product will destroy the economy.” 😄

‘Killer robots’ must be banned, UN secretary-general says

A global workspace in language models: Anthropic researchers detail J-space, a small set of neural patterns in Claude that reveals internal thoughts that don’t appear in the model’s output.

New models and services

OpenAI launches GPT-Live voice model series ahead of broad GPT-5.6 release

OpenAI debuts ChatGPT Work, an agentic tool for automating business workflows

SpaceXAI’s newest AI model Grok 4.5 dramatically undercuts Anthropic and OpenAI on price

Meta launches flagship Muse Spark 1.1 model with multi-agent upgrades

Meta launches image generation model with coding, search capabilities And in the process, pissed off just about everybody. (You can turn most of it off.) Update: Welp, that didn’t last long.

Anthropic brings Cowork out of the desktop and onto web and mobile

Microsoft is reportedly ditching OpenAI’s and Anthropic’s AI models in favor of its own to cut costs

Ex-GitHub chief’s Entire opens distributed Git network for the AI agent era

Salesforce enhances Slackbot with connectors to the entire platform ecosystem

Superhuman launches Docs, merging writing, AI and data for document collaboration

Refiant goes where rivals only promised with a 10 million-token AI model

Mindbeam sets generative AI models to task on drug design, hunting for better pain meds

DeepFabric ships more than 50 AI agents for supply chain operations

Exclusive: Corvic AI launches V5 to turn one-off prompts into repeatable workflows

Amazon’s Mechanical Turk service now on life support as it stops accepting new users

Selected exclusive coverage of the RAISE Summit

The highlights from Intellyx’s Jason Bloomberg: Pulling out the AI stops at RAISE Summit Paris

Cerebras Systems positions inference speed as the defining edge in AI infrastructure

VAST Data targets KV cache storage and neoclouds as AI infrastructure enters the exabyte era

AMD targets system-level AI infrastructure optimization as agentic workloads reshape enterprise compute

Fast token generation emerges as the key differentiator as heterogeneous inference takes hold

Policy

Apple sues OpenAI, former staffers over alleged intellectual property theft

Money matters

PitchBook: US venture funding hits $412.7B in first half as AI deals dominate

Mercor buys Deeptune to build training environments for AI agents

Open-source AI model developer MiniMax raises $2B in funding

Swedish vibe-coding startup Lovable reportedly in talks to raise $300M at a valuation of $13.2B

Prime Intellect raises $130M at $1B valuation for its AI training platform

Web data scraping infrastructure startup Oxylabs reels in $130M in its first funding round

Norm Ai nabs $120M at $1.2B valuation to bring AI agents to the law

AI agent startup Lyzr reportedly raising $100M at $500M valuation

Market data platform startup Databento closes $97M round after drawing $300M in investor demand

Open-source AI developer tool Ollama raises $65M to grow its platform

Kaon AI raises $60M to build the next generation of AI-driven interactive entertainment

AI post-training startup Bespoke Labs raises $40M in funding

With $25M in funding, Alta AI aims to accelerate go-to-market automation

Tangos grabs $20M in funding to take on the bad guys by automating financial crime investigations

Around the enterprise: The new chip gold rush

Money matters

Memory chipmaker SK hynix seeks to raise $28B through US IPO

And the results: Memory chip giant SK hynix bags $26.5B in blockbuster US listing Shares opened Friday morning 14% above their asking price, ending the day up 13%.

Apple promises to buy $30B worth of US-made chips from Broadcom

Amazon launches $25B bond sale to fund AI infrastructure

Anthropic inks $19B AI data center lease with TeraWulf

Inference chip startup SambaNova valued at $11B in $1B funding round

Samsung says memory chip demand will drive 19-fold jump in operating profit

Nvidia-backed startup Nscale locks in $900M for data-center buildout

Syntiant, which makes low-power chips for on-device AI, files for initial public offering

Infoblox plans to acquire Kentik in DNS and network observability deal (per CRN)

Germany’s QuantumDiamonds closed a €91 million funding round to scale production of its quantum-based semiconductor inspection technology

Arkenstone Defense launches with $35M to help startups sell to the Pentagon

New products and services

IBM debuts compact z17 mainframes and LinuxONE servers for on-premises enterprise AI

Report: China’s DeepSeek follows OpenAI in developing its own custom inference chips

Policy

EU finds that Meta breached bloc’s rules with its social network interfaces

UK to regulate cloud service providers Microsoft, Google and others to protect financial stability

Cyber beat: The first auto-ransomware attack