Artificial Intelligence Technology's Week Beyond the Models

The model launches took the headlines this week; the rest of the news cycle is where the artificial intelligence technology story actually moved. Our feature this week goes deep on the pricing-and-access turbulence on the models themselves. This roundup is everything else the same week turned up, from the capital funding these labs to the labor market now absorbing their tools.

---

Menlo Ventures Bets the Firm Again With a Record $3 Billion Fund

On June 23, on its 50th anniversary, Menlo Ventures closed the largest fund in its history: $3 billion, split across Menlo Ventures XVII for seed and Series A deals and Menlo Inflection IV for Series B and later-stage growth. The firm framed the raise, plainly, as going all in on AI.

The confidence has a paper trail. Menlo first backed Anthropic in 2023, when the company was pre-product and pre-revenue, then wrote the largest check in its history to lead Anthropic's Series D and followed in every round since. That position, built on roughly $1 billion invested, is now worth close to $14 billion against an Anthropic valuation north of $900 billion. Managing partner Shawn Carolan has called it a bet-the-firm moment.

For anyone shipping software in technology and telecommunications, the artificial intelligence technology companies you will build on, buy from, or compete with next is being funded now, at a scale that makes the downstream model launches look like the second act. TechCrunch and Bloomberg both covered the raise.

Source: TechCrunch

---

Anthropic Lines Up an October IPO That Could Clear a Trillion Dollars

Anthropic is not staying private to spend that kind of money. The company confidentially filed for an IPO with the SEC on June 1, targeting an October 2026 Nasdaq listing with Goldman Sachs, JPMorgan, and Morgan Stanley leading an offering reported to seek more than $60 billion.

The figures around it are large even by this market's standards. Anthropic's last private mark was a $65 billion Series H in May at a $965 billion post-money valuation, and secondary markets were pricing an implied $1.05 trillion to $1.15 trillion ahead of any listing. No share price, share count, or listing date is set, and the company has said timing depends on market conditions.

For anyone building on the models, a big shift here will be disclosure. A frontier lab on public markets has to publish audited revenue and the cost of the compute behind its models, which would give teams their first hard look at the unit economics under the tools they rent by the token. Teams that already meter their own runs will read those filings fastest; tracking token spend against tool spend is the discipline AgentPMT builds in by itemizing per-token and per-tool-call cost on every run, so the economics are legible before the S-1, not after. Futurum Group and the wider financial press reported the filing.

Source: Futurum Group

---

Google's Gemini 3.5 Pro Slips to July as Top Researchers Walk

Google spent the week on the back foot. It pushed Gemini 3.5 Pro's general availability from June into July, leaving the model in limited Vertex AI enterprise preview after Sundar Pichai told the I/O audience on May 19 to "give us until next month."

The delay arrived with an exodus of names. Noam Shazeer, a vice president of engineering on Gemini and a co-author of the 2017 "Attention Is All You Need" paper that introduced the transformer, is leaving for OpenAI. John Jumper, the DeepMind scientist whose AlphaFold work shared the 2024 Nobel Prize in Chemistry, is leaving for Anthropic. Four senior Gemini researchers announced departures within days, and Alphabet shares dropped about 5% on June 22, erasing roughly $225 billion in market value in a single session.

Read past the drama and it is a competition story, and competition is the builder's friend. A third serious frontier lab pushing to ship strengthens the fallback bench for anyone who stays model-agnostic, so a slip at one artificial intelligence technology company becomes a routing decision rather than a roadmap problem. That is the posture AgentPMT is built around: a model gateway routes every call so the underlying model swaps in and out of a workflow without a rebuild, so twelve frontier models sitting less than a point apart becomes a reason to build on the orchestration layer rather than bet on a single lab. TechTimes and Bind AI both tracked the delay and the departures.

Source: TechTimes

---

Anthropic Opens Claude Science and $30,000 Research Grants

On July 1 Anthropic put its own model to work on research. Claude Science is a workbench that runs multi-agent orchestration across more than 60 scientific databases on Claude Opus 4.8, open to paid subscribers with no enterprise gating and able to run on a lab's own infrastructure.

It comes with money attached. The AI for Science program is funding up to 50 projects with as much as $30,000 in API credits each, and Modal is adding up to $2,000 in compute for selected work. Applications close July 15, awards are announced by July 31, and funded projects run September 1 to December 1, with an early focus on biomedical, graduate, and postdoctoral research.

This is the most build-forward item of the week: frontier compute pointed at real science and handed to researchers rather than gated behind procurement. It is also a preview of how AI technology companies seed the work that turns into next year's products and tools. TechTimes covered the launch alongside Anthropic's own program pages.

Source: TechTimes

---

The June Jobs Report Puts AI in the Frame

The month's macro data gave the model news a sharper backdrop. U.S. employers added just 57,000 jobs in June, roughly half the consensus estimate, while the unemployment rate slipped to 4.2%, largely because labor-force participation fell to 61.5%, its lowest reading since March 2021. Leisure and hospitality lost 61,000 positions; professional and business services added 36,000.

AI showed up in the analysis, not just the headline. Goldman Sachs economists estimated that AI is already pulling 10,000 to 15,000 jobs a month out of payroll growth, concentrated in tech, management consulting, and design, and projected it could displace around 15 million U.S. workers, roughly 9% of the workforce, over the next decade. A June GMAC survey found one in three employers has replaced an entry-level role with AI rather than hiring for it.

Automation is now a measurable force in the labor data across technology and telecommunications, and the teams that come out ahead are the ones that stay exact about which tasks it genuinely absorbs and where a person still sets the direction. CNBC published the report; the labor-impact estimates are Goldman Sachs's.

Source: CNBC

---

Step back and no one model captures the week. In one week the money (a record venture fund and a filing aimed at a trillion-dollar debut), the talent (a Nobel laureate and a transformer co-author changing badges), and the labor market all repriced around artificial intelligence technology at once. For anyone building on top of it, that argues for the same posture the model news did: stay portable across vendors, keep the economics of every run in view, and hold a human hand on the calls automation cannot make yet. Build for a market that reprices this often, and a week like this reads as a dashboard update instead of a fire drill.

---

Sources

• After Betting the Firm on Anthropic, Menlo Ventures Raises a Victorious $3B Fund, TechCrunch
• Anthropic Backer Menlo Ventures Lands $3 Billion in Its Largest-Ever Haul, Bloomberg
• Anthropic Files for IPO, Looking to Beat OpenAI to the Punch, Futurum Group
• Gemini 3.5 Pro Slips to July and Four Senior Google Researchers Just Left, Bind AI
• Gemini 3.5 Pro Cleared for July Launch, TechTimes
• Anthropic Launches Claude Science: AI Research Workbench Open to All Paid Subscribers, TechTimes
• Jobs Report June 2026: Payrolls Rise 57,000, CNBC
• June Jobs Report: 57,000 Payrolls Miss Consensus by Half, TechTimes
• AI News Today, July 3, 2026, Build Fast with AI