1
00:00:04,550 --> 00:00:07,349
Robinhood lets AI agents trade stocks and spend.

2
00:00:07,670 --> 00:00:10,779
Third party agents now hold the keys to retail order flow,

3
00:00:10,939 --> 00:00:12,229
and the first ones to clear will

4
00:00:12,229 --> 00:00:14,770
route everything through the lowest latency broker.

5
00:00:14,989 --> 00:00:17,989
The MCP integration turns every spending limit into

6
00:00:17,989 --> 00:00:19,979
a live parameter an agent can optimize.

7
00:00:20,069 --> 00:00:21,430
Against market data.

8
00:00:21,780 --> 00:00:25,530
Banks without equivalent hooks lose the next layer of automated volume.

9
00:00:25,760 --> 00:00:29,840
Expect the first wave of agent-driven credit lines to surface within 9 months,

10
00:00:30,120 --> 00:00:32,799
forcing legacy platforms to either expose their own

11
00:00:32,799 --> 00:00:35,580
protocols or seed the entire mid-tier book.

12
00:00:36,200 --> 00:00:38,889
Snowflake signed $6 billion AWS deal.

13
00:00:39,169 --> 00:00:40,470
Shares surged sharply.

14
00:00:40,729 --> 00:00:43,849
Choosing Graviton over standard instances cuts Snowflake's

15
00:00:43,849 --> 00:00:45,909
own compute spend by roughly a third

16
00:00:46,029 --> 00:00:47,349
once the migration finishes.

17
00:00:47,849 --> 00:00:50,799
That margin lift is already baked into the raised outlook.

18
00:00:51,115 --> 00:00:53,814
It also locks the company deeper into AWS

19
00:00:53,814 --> 00:00:56,465
tooling that smaller rivals cannot match at scale.

20
00:00:56,755 --> 00:00:59,115
Watch the next three enterprise RFPs.

21
00:00:59,275 --> 00:01:03,354
Any vendor still quoting X86 pricing will either discount aggressively

22
00:01:03,354 --> 00:01:06,095
or lose the deal to the new Graviton-backed rates.

23
00:01:06,860 --> 00:01:11,559
Cognition raises over $1 billion at $26 billion valuation for Devin.

24
00:01:11,900 --> 00:01:15,099
Every Series C at this multiple now sets the benchmark for

25
00:01:15,099 --> 00:01:18,620
what an autonomous coding team must deliver inside 12 months.

26
00:01:18,900 --> 00:01:21,339
Devin's backers are pricing and replacement of

27
00:01:21,339 --> 00:01:23,680
entire junior engineering benches at scale.

28
00:01:23,970 --> 00:01:26,470
The pressure hits GitHub and JetBrains first,

29
00:01:26,620 --> 00:01:29,019
since their plug-in ecosystems become the obvious

30
00:01:29,019 --> 00:01:31,059
integration layer for any new agent.

31
00:01:31,400 --> 00:01:34,300
That leaves every incumbent tool vendor with one choice,

32
00:01:34,480 --> 00:01:37,300
embed similar agent loops before their renewal cycle

33
00:01:37,580 --> 00:01:40,889
or watch ARR migrate to the new default environment.

34
00:01:41,349 --> 00:01:45,419
Base 10I's $1 billion raise at $11 billion valuation.

35
00:01:45,819 --> 00:01:48,949
Enterprise models serving margins compress the moment inference

36
00:01:48,949 --> 00:01:51,779
moves behind an SLA with uptime guarantees.

37
00:01:52,110 --> 00:01:54,709
Base 10's Ra assumes it can capture that layer

38
00:01:54,709 --> 00:01:57,629
before the largest customers finish their own hardware builds.

39
00:01:57,910 --> 00:02:01,029
The first sign will be when two large fintechs announce internal

40
00:02:01,029 --> 00:02:04,610
clusters that match Base 10's claimed latency at half the cost.

41
00:02:04,940 --> 00:02:08,720
Any delay lets the next batch of on-prem optimizations land first,

42
00:02:08,869 --> 00:02:13,350
and the valuation math flips from growth to replacement cost inside 18 months.

43
00:02:13,919 --> 00:02:16,470
Mistral launches Vibe remote coding agents.

44
00:02:16,720 --> 00:02:20,139
Long running sessions change the economics of remote dev work.

45
00:02:20,320 --> 00:02:22,479
A single agent can iterate through an entire

46
00:02:22,479 --> 00:02:25,750
feature branch overnight without burning local GPU cycles.

47
00:02:26,080 --> 00:02:28,720
The work mode toggle keeps context across days,

48
00:02:28,880 --> 00:02:29,740
not hours.

49
00:02:29,960 --> 00:02:33,649
Expect GitHub to cut co-pilot prices within 6 months or lose

50
00:02:33,649 --> 00:02:36,860
the async segment to teams that value handoff over chat.

51
00:02:37,100 --> 00:02:39,899
Smaller consultancies gain an edge because they no longer

52
00:02:39,899 --> 00:02:42,360
need round the clock staff on every project.

53
00:02:42,699 --> 00:02:45,809
Margin pressure on local tools appears in renewal rates.

54
00:02:46,179 --> 00:02:50,130
Anthropic releases clawed Opus 4.8 with agentic tools.

55
00:02:50,460 --> 00:02:53,860
Sub-agent swarms are reliable enough now for production workflows

56
00:02:53,860 --> 00:02:56,880
that run without constant human oversight on every step.

57
00:02:57,100 --> 00:03:00,259
Improved honesty at the same price point shifts the cost of air.

58
00:03:00,440 --> 00:03:03,479
Correction away from customers and onto the model provider.

59
00:03:03,880 --> 00:03:07,679
OpenAI must now ship comparable swarm tooling by year end

60
00:03:07,679 --> 00:03:11,190
or watch its coding benchmarks lose relevance in large deployments.

61
00:03:11,479 --> 00:03:13,679
The margin pressure on API calls hits

62
00:03:13,679 --> 00:03:16,820
when swarm volume scales under Enterprise SLAs.

63
00:03:17,240 --> 00:03:21,699
KPMG deploys Cloud to 276,000 employees globally.

64
00:03:22,039 --> 00:03:23,600
Junior consultants now handle task.

65
00:03:23,750 --> 00:03:27,369
Tasks that used to require manager review within the first week of onboarding.

66
00:03:27,699 --> 00:03:30,699
Tax and private equity workflows see the fastest lift

67
00:03:30,699 --> 00:03:33,850
because those practices already run on templated data flows.

68
00:03:34,139 --> 00:03:37,419
Deloitte has to match the deployment scale inside 18 months or

69
00:03:37,419 --> 00:03:41,009
risk losing bids where clients demand AI native delivery teams.

70
00:03:41,380 --> 00:03:43,899
Rate cards for entry level work will face compression

71
00:03:43,899 --> 00:03:46,899
once the managed agents prove stable in client audits.

72
00:03:47,550 --> 00:03:51,979
OpenAI Deploy Co $4 billion consulting subsidiary advances.

73
00:03:52,309 --> 00:03:54,789
Forward deployed engineers changed the sales cycle

74
00:03:54,789 --> 00:03:57,509
from model access to full outcome guarantees.

75
00:03:57,669 --> 00:04:02,210
Accenture must either acquire comparable AI deployment talent inside 12 months

76
00:04:02,389 --> 00:04:06,750
or accept lower margins on projects that now require model fine tuning on site.

77
00:04:06,990 --> 00:04:09,830
The trajectory favors labs that control both the model

78
00:04:09,830 --> 00:04:12,740
and the integration layer over pure strategy firms.

79
00:04:13,070 --> 00:04:16,589
Mid-tier integrators without any lab backing will see their pipeline

80
00:04:16,589 --> 00:04:19,709
stall within the next two quarters as clients shift spend.

81
00:04:20,218 --> 00:04:22,759
Meta rolls out paid chatbot subscriptions.

82
00:04:22,979 --> 00:04:25,098
Every new paid user still adds to the

83
00:04:25,098 --> 00:04:27,889
same compute bill that free accounts already strain.

84
00:04:28,178 --> 00:04:30,968
Meta's move locks higher usage behind paywalls,

85
00:04:30,979 --> 00:04:32,838
while free tiers degrade in speed.

86
00:04:33,088 --> 00:04:35,859
This forces OpenAI to either match the price point

87
00:04:35,859 --> 00:04:39,739
or pivot to enterprise-only contracts before mid-market customers defect.

88
00:04:40,089 --> 00:04:44,070
The buried data licensing clauses will decide whether smaller agents can train

89
00:04:44,070 --> 00:04:47,720
on paid conversations or get locked out by Q3 next year,

90
00:04:47,970 --> 00:04:50,790
cutting off access to conversation data entirely.

91
00:04:51,130 --> 00:04:55,089
Snowflake Cortex adds Gemini 3.5 flash multimodal.

92
00:04:55,369 --> 00:04:57,799
Latency improvements matter less than who absorbs

93
00:04:57,799 --> 00:05:00,359
the variable cost when query volume spikes.

94
00:05:00,609 --> 00:05:01,850
Agent-driven video queries.

95
00:05:02,119 --> 00:05:06,549
Inside existing dashboards now hit production SLAs without custom pipelines.

96
00:05:06,869 --> 00:05:10,279
The real pressure lands on Snowflake's compute margins once Gemini's

97
00:05:10,279 --> 00:05:13,160
token pricing layers on top of every warehouse query.

98
00:05:13,470 --> 00:05:16,790
This forces data bricks to bundle their own multimodal models at a

99
00:05:16,790 --> 00:05:21,100
flat rate or watch customers migrate entire workflows by early next year

100
00:05:21,320 --> 00:05:23,500
without losing margin share on the base platform.

101
00:05:24,100 --> 00:05:28,299
Cerebras eyed as top 2026 AI chip IPO candidate.

102
00:05:28,579 --> 00:05:32,100
Performance claims on wafer scale chips still hinge on sustained

103
00:05:32,100 --> 00:05:35,260
yields that few foundries have proven at commercial volumes.

104
00:05:35,500 --> 00:05:38,880
An IPO next year would value cerebras against that risk,

105
00:05:39,100 --> 00:05:40,540
not just benchmark numbers.

106
00:05:40,820 --> 00:05:44,299
Investors will price in the OpenAI partnership only after the

107
00:05:44,299 --> 00:05:47,779
1st 10,000 wafer run clears defect rates below 2%.

108
00:05:48,130 --> 00:05:51,950
This forces Nvidia to accelerate its own custom silicon roadmap

109
00:05:52,130 --> 00:05:55,290
or lose the account entirely once inference clusters scale

110
00:05:55,290 --> 00:05:57,679
beyond current supply by the end of next year.

111
00:05:58,089 --> 00:06:01,570
Canada issues privacy ruling on chat GPT training data.

112
00:06:01,890 --> 00:06:04,730
Deletion orders on scraped data sets create a precedent

113
00:06:04,730 --> 00:06:07,829
that hits every frontier lab using public web crawls,

114
00:06:08,209 --> 00:06:09,829
including those still in stealth.

115
00:06:10,010 --> 00:06:14,890
OpenAI must now either license equivalent data at much higher cost or restrict.

116
00:06:15,394 --> 00:06:17,445
updates for Canadian users entirely.

117
00:06:17,744 --> 00:06:20,945
Expect the same regulators to demand audit logs on every

118
00:06:20,945 --> 00:06:24,154
training token within 18 months or face daily fines,

119
00:06:24,225 --> 00:06:27,144
raising compliance overhead across the industry as smaller

120
00:06:27,144 --> 00:06:30,045
labs without dedicated legal teams exit first,

121
00:06:30,184 --> 00:06:32,084
when enforcement begins next quarter.