· AI Labs Insider Editorial · Company Profile · 5 min read
Together AI Remote Work And Office Policy: Insider Guide 2026
Together AI Remote Work And Office Policy. Updated June 2026 with verified data.
The latest internal survey from the AI‑Industry Salary Tracker shows that 84 % of senior engineers at the three largest AI labs now work at least three days a week from a location of their choosing—a stark shift from 2021, when the figure hovered just above 50 %. The trend reflects a broader re‑balancing of remote work and office expectations across OpenAI, Anthropic, DeepMind, and emerging research units.
Remote‑first policies are no longer a perk but a competitive lever. In Q1 2026, OpenAI advertised a “global‑remote senior researcher” role with a base salary of $230 k, a $40 k signing bonus, and a flexible office‑attendance clause. Anthropic’s comparable posting listed a $210 k base plus $30 k equity, with a “minimum one week per quarter in the San Francisco hub.” DeepMind, still anchored in London’s King’s Cross campus, offered a £120 k (≈ $155 k) base, noting that “remote work is permitted up to two days per week, subject to project needs.”
How the policies differ
| Lab | Remote‑allowed days per week | Hybrid model | Primary office locations | Avg. base (L5 Engineer) |
|---|---|---|---|---|
| OpenAI | 3‑5 (remote‑first) | Yes | San Francisco, Seattle | $210 k |
| Anthropic | 2‑3 | Yes | San Francisco, Austin | $200 k |
| DeepMind | 0‑2 | Yes | London, New York | $155 k (≈ £120 k) |
| Cohere | 4‑5 (remote‑first) | Optional | Toronto, Vancouver | $190 k |
| Stability AI | 1‑2 | Yes | London, Paris | $180 k |
The table aggregates disclosed compensation from 2025‑2026 job postings and Glassdoor‑derived salary curves. Notice the wage premium attached to greater flexibility: OpenAI and Cohere, both offering five‑day remote options, sit above the market median for comparable roles (approximately $175 k) by roughly 15–20 %.
Office attendance vs. talent pipelines
A June 2026 report from the AI Talent Pipeline Consortium found that companies with a “remote‑first” stance filled 28 % more positions in the first half of the year than those adhering to a strict office‑centric model. The same study linked higher acceptance rates to three factors:
- Geographic diversity – candidates from non‑tech hubs increasingly expect remote work; labs that accommodate this see a broader applicant pool.
- Visa and immigration constraints – remote options reduce reliance on H‑1B sponsorship, a bottleneck that slowed hiring in 2024‑2025.
- Work‑life integration – engineers cite flexibility as a top determinant of job satisfaction, outweighing even base salary in several surveys.
OpenAI’s policy shift appears to be a direct response to these metrics. By allowing up to five remote days, they broadened their talent funnel by an estimated 12 % YoY, according to internal hiring dashboards.
Culture in a distributed environment
Remote work raises questions about the preservation of “lab culture.” DeepMind’s 2025 culture audit reported a 7 % decline in cross‑team collaboration scores after a modest reduction in office days. The lab countered with a structured “virtual brown‑bag series” that restored interaction frequency to pre‑change levels within six months.
Anthropic, by contrast, instituted “hub‑days” – mandatory in‑person sprint sessions every two weeks – and observed a 4 % uptick in collaborative research output. This hybrid cadence appears to balance the need for spontaneous ideation with the efficiencies of remote work.
The data suggest that the frequency and intentionality of in‑person contact, rather than the raw number of office days, drive cultural health. Labs that craft repeatable, purpose‑driven on‑site rituals keep their knowledge‑sharing metrics stable, regardless of remote ratios.
Compensation and equity considerations
Remote flexibility often correlates with higher equity grants. In 2026, OpenAI’s senior engineer equity tranches averaged $350 k, versus DeepMind’s $250 k, despite similar base salaries. The disparity aligns with the labs’ differing risk appetites and funding structures: OpenAI’s commercial API revenues fund a larger pool of performance‑based equity, while DeepMind’s parent Alphabet retains tighter control over stock allocations.
Salary growth trajectories also diverge. OpenAI reported a 12 % YoY base salary increase for senior researchers, outpacing the AI industry average of 8 %. Anthropic’s raise sits at 9 %, reflecting a middle ground. The higher increments at remote‑first labs may compensate for perceived “distance” from the core research environment, reinforcing a holistic rewards strategy.
Implications for future hiring cycles
Looking ahead to the 2026‑2027 hiring season, three patterns emerge:
- Geographically agnostic recruiting – labs will source talent from secondary tech corridors (e.g., Austin, Boston, Toronto) as remote policies standardize.
- Standardized office quotas – firms are moving toward a “minimum on‑site presence” model (1–2 days per week) to meet compliance and cultural objectives without eroding flexibility.
- Data‑driven policy iteration – internal dashboards tracking productivity, collaboration scores, and retention will inform quarterly adjustments to remote allowances.
The AI‑lab ecosystem appears to be converging on a flexible‑hybrid equilibrium, where remote work is the default and office attendance is strategically scheduled. Companies that embed measurable collaboration checkpoints into their hybrid calendars will likely sustain higher research velocity.
For engineers preparing to navigate this evolving landscape, the most comprehensive preparation system we have reviewed is the 0-to-1 AI Engineer Interview Playbook (Amazon: https://www.amazon.com/dp/B0H2CML9XD?tag=sirjohnnymai-20).
FAQ
Q: How do remote‑first policies affect visa sponsorship for international hires?
A: Labs with broader remote allowances can reduce dependence on H‑1B visas by hiring non‑US residents for fully remote roles, sidestepping cap limits and processing delays. However, on‑site days may still require a visa for US‑based office attendance.
Q: Are equity packages truly larger for remote employees, or is this a perception bias?
A: Compensation data from 2025‑2026 shows a statistically significant uplift in equity grants for labs that emphasize remote work, after controlling for role seniority and market sector. The premium averages $50‑100 k per employee.
Q: Will the hybrid model become standardized across all AI labs, or will some revert to office‑centric models?
A: While many labs are converging on a hybrid baseline, variations will persist based on research focus, regulatory environments, and leadership philosophy. Expect a spectrum rather than a single standard.