From what I can tell, either way you accent this word is probably fine.
My understanding, based on Biblical grammar
My understanding is that the accent in this case goes on the 'mo' syllable1, due to the rule of "nasog achor."
This rule says that when multi-syllabic Word A is followed (without disjunctive cantillation) by Word B, and Word B has an accent on its first (or only) syllable, then Word A's accent goes to its penultimate syllable, even if it would normally be on the last syllable. (See, for example, Exodus 20:19, eh-LO-hei CHE-sef vei-lo-HEI za-HAV. Not eh-lo-HEI CHE-sef, as that would put stressed syllables back-to-back.)
So, in the first verse you cite, Psalms 104:14
לְה֥וֹצִיא לֶ֝֗חֶם מִן־הָאָֽרֶץ׃ ...
Because the cantillation on "lehotzi" is conjuctive and the accent on "lechem" is on the first syllable, the accent moves from the last syllable, 'tzi', to the penultimate one, 'ho'.
(I'm not sure why this rule doesn't apply to the second verse you cite, since "lecha" has only one syllable. Maybe someone can edit this to explain.)
Similarly, in the blessing, the word "lechem" has an accent on the first syllable, so it pushes the accent back one spot on the previous word, "hamotzi."
Survey of Siddurim
I (and other yodeyans) looked in various different siddurim, all of which have marks to indicate when the accent is on the penultimate syllable, for printed confirmation of my understanding. Unfortunately, I found that two confirmed my observation by placing an accent mark on the 'mo' syllable, while two did not. Perhaps someone can provide reasoning or documentation to explain the position of the two that did not.
Siddurim that confirm my understanding and place the accent on the previous syllable in this case:
Siddurim that defy my understanding in this case:
Explanation from Koren
I wrote to Koren Publishers' Senior Hebrew Grammar Expert, Efrat Gross, asking why they don't print the accent on the penultimate syllable in this case. She replied with this answer from the Koren Siddur's chief proofreader, Chanan Ariel:
כפי שנאמר בדברי ההסבר, מיעטנו ב"נסוג אחור" מפני שלא ברור אם יש להחיל את כללי לשון המקרא על לשון התפילה בעניין הזה. במקום שלא מסומן בו "נסוג אחור" אין פירושו של דבר שהקורא בנסיגת הטעם איננו צודק, אלא שחשבנו שאין חובה לקרוא בנסיגת הטעם. מאחר שיש כאן שינוי מלשון הפסוק שיש בו צורת מקור ולא צורת בינוני, חשבנו שעדיף שלא לקבוע מסמרות בעניין מקום הטעם.
Here's my translation (with some help from Shalom and Avi Shmidman):
As is stated in the explanatory notes, we limited [application of] "Nasog Achor" because it's not clear if the rules of Biblical language apply to the language of prayer in this matter. Where "*Nasog Achor" is not indicated, this does not mean that one who reads with the accent pulled back isn't correct, but that we thought that there's no obligation to read with the accent pulled back. Since this text differs from the language of the [Biblical] verse, which is in the infinitive form2, rather the participle form3, we thought that it would be better not nail down the placement of the accent.
So, at least according to Koren, either way is acceptable.
1. Throughout this answer, I'm using the word "syllable" to refer to the Hebrew grammar concept of tenua, which is a little different from what "syllable" generally means. A tenua always contains a non-shva vowel and may also include adjacent letters with shvas under them. A shva never forms a tenua on its own; it's always assigned to the tenua formed by an adjacent letter.
2. "להוציא" - "to extract"
3. "המוציא" - "The One Who extracts"